20220524.md

ArXiv cs.CV --Tue, 24 May 2022

1.VQA-GNN: Reasoning with Multimodal Semantic Graph for Visual Question Answering ⬇️

Visual understanding requires seamless integration between recognition and reasoning: beyond image-level recognition (e.g., detecting objects), systems must perform concept-level reasoning (e.g., inferring the context of objects and intents of people). However, existing methods only model the image-level features, and do not ground them and reason with background concepts such as knowledge graphs (KGs). In this work, we propose a novel visual question answering method, VQA-GNN, which unifies the image-level information and conceptual knowledge to perform joint reasoning of the scene. Specifically, given a question-image pair, we build a scene graph from the image, retrieve a relevant linguistic subgraph from ConceptNet and visual subgraph from VisualGenome, and unify these three graphs and the question into one joint graph, multimodal semantic graph. Our VQA-GNN then learns to aggregate messages and reason across different modalities captured by the multimodal semantic graph. In the evaluation on the VCR task, our method outperforms the previous scene graph-based Trans-VL models by over 4%, and VQA-GNN-Large, our model that fuses a Trans-VL further improves the state of the art by 2%, attaining the top of the VCR leaderboard at the time of submission. This result suggests the efficacy of our model in performing conceptual reasoning beyond image-level recognition for visual understanding. Finally, we demonstrate that our model is the first work to provide interpretability across visual and textual knowledge domains for the VQA task.

2.Flexible Diffusion Modeling of Long Videos ⬇️

We present a framework for video modeling based on denoising diffusion probabilistic models that produces long-duration video completions in a variety of realistic environments. We introduce a generative model that can at test-time sample any arbitrary subset of video frames conditioned on any other subset and present an architecture adapted for this purpose. Doing so allows us to efficiently compare and optimize a variety of schedules for the order in which frames in a long video are sampled and use selective sparse and long-range conditioning on previously sampled frames. We demonstrate improved video modeling over prior work on a number of datasets and sample temporally coherent videos over 25 minutes in length. We additionally release a new video modeling dataset and semantically meaningful metrics based on videos generated in the CARLA self-driving car simulator.

3.Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding ⬇️

We present Imagen, a text-to-image diffusion model with an unprecedented degree of photorealism and a deep level of language understanding. Imagen builds on the power of large transformer language models in understanding text and hinges on the strength of diffusion models in high-fidelity image generation. Our key discovery is that generic large language models (e.g. T5), pretrained on text-only corpora, are surprisingly effective at encoding text for image synthesis: increasing the size of the language model in Imagen boosts both sample fidelity and image-text alignment much more than increasing the size of the image diffusion model. Imagen achieves a new state-of-the-art FID score of 7.27 on the COCO dataset, without ever training on COCO, and human raters find Imagen samples to be on par with the COCO data itself in image-text alignment. To assess text-to-image models in greater depth, we introduce DrawBench, a comprehensive and challenging benchmark for text-to-image models. With DrawBench, we compare Imagen with recent methods including VQ-GAN+CLIP, Latent Diffusion Models, and DALL-E 2, and find that human raters prefer Imagen over other models in side-by-side comparisons, both in terms of sample quality and image-text alignment. See https://imagen.research.google/ for an overview of the results.

4.Exposing Outlier Exposure: What Can Be Learned From Few, One, and Zero Outlier Images ⬇️

Traditionally anomaly detection (AD) is treated as an unsupervised problem utilizing only normal samples due to the intractability of characterizing everything that looks unlike the normal data. However, it has recently been found that unsupervised image anomaly detection can be drastically improved through the utilization of huge corpora of random images to represent anomalousness; a technique which is known as Outlier Exposure. In this paper we show that specialized AD learning methods seem actually superfluous and huge corpora of data expendable. For a common AD benchmark on ImageNet, standard classifiers and semi-supervised one-class methods trained to discern between normal samples and just a few random natural images are able to outperform the current state of the art in deep AD, and only one useful outlier sample is sufficient to perform competitively. We investigate this phenomenon and reveal that one-class methods are more robust towards the particular choice of training outliers. Furthermore, we find that a simple classifier based on representations from CLIP, a recent foundation model, achieves state-of-the-art results on CIFAR-10 and also outperforms all previous AD methods on ImageNet without any training samples (i.e., in a zero-shot setting).

5.Graph-theoretical approach to robust 3D normal extraction of LiDAR data ⬇️

Low dimensional primitive feature extraction from LiDAR point clouds (such as planes) forms the basis of majority of LiDAR data processing tasks. A major challenge in LiDAR data analysis arises from the irregular nature of LiDAR data that forces practitioners to either regularize the data using some form of gridding or utilize a triangular mesh such as triangulated irregular network (TIN). While there have been a handful applications using LiDAR data as a connected graph, a principled treatment of utilizing graph-theoretical approach for LiDAR data modelling is still lacking. In this paper, we try to bridge this gap by utilizing graphical approach for normal estimation from LiDAR point clouds. We formulate the normal estimation problem in an optimization framework, where we find the corresponding normal vector for each LiDAR point by utilizing its nearest neighbors and simultaneously enforcing a graph smoothness assumption based on point samples. This is a non-linear constrained convex optimization problem which can then be solved using projected conjugate gradient descent to yield an unique solution. As an enhancement to our optimization problem, we also provide different weighted solutions based on the dot product of the normals and Euclidean distance between the points. In order to assess the performance of our proposed normal extraction method and weighting strategies, we first provide a detailed analysis on repeated randomly generated datasets with four different noise levels and four different tuning parameters. Finally, we benchmark our proposed method against existing state-of-the-art approaches on a large scale synthetic plane extraction dataset. The code for the proposed approach along with the simulations and benchmarking is available at this https URL.

6.SiPRNet: End-to-End Learning for Single-Shot Phase Retrieval ⬇️

Traditional optimization algorithms have been developed to deal with the phase retrieval problem. However, multiple measurements with different random or non-random masks are needed for giving a satisfactory performance. This brings a burden to the implementation of the algorithms in practical systems. Even worse, expensive optical devices are required to implement the optical masks. Recently, deep learning, especially convolutional neural networks (CNN), has played important roles in various image reconstruction tasks. However, traditional CNN structure fails to reconstruct the original images from their Fourier measurements because of tremendous domain discrepancy. In this paper, we design a novel CNN structure, named SiPRNet, to recover a signal from a single Fourier intensity measurement. To effectively utilize the spectral information of the measurements, we propose a new Multi-Layer Perception block embedded with the dropout layer to extract the global representations. Two Up-sampling and Reconstruction blocks with self-attention are utilized to recover the signals from the extracted features. Extensive evaluations of the proposed model are performed using different testing datasets on both simulation and optical experimentation platforms. The results demonstrate that the proposed approach consistently outperforms other CNN-based and traditional optimization-based methods in single-shot maskless phase retrieval. The source codes of the proposed method have been released on Github: this https URL.

7.Decoder Denoising Pretraining for Semantic Segmentation ⬇️

Semantic segmentation labels are expensive and time consuming to acquire. Hence, pretraining is commonly used to improve the label-efficiency of segmentation models. Typically, the encoder of a segmentation model is pretrained as a classifier and the decoder is randomly initialized. Here, we argue that random initialization of the decoder can be suboptimal, especially when few labeled examples are available. We propose a decoder pretraining approach based on denoising, which can be combined with supervised pretraining of the encoder. We find that decoder denoising pretraining on the ImageNet dataset strongly outperforms encoder-only supervised pretraining. Despite its simplicity, decoder denoising pretraining achieves state-of-the-art results on label-efficient semantic segmentation and offers considerable gains on the Cityscapes, Pascal Context, and ADE20K datasets.

8.LILA-BOTI : Leveraging Isolated Letter Accumulations By Ordering Teacher Insights for Bangla Handwriting Recognition ⬇️

Word-level handwritten optical character recognition (OCR) remains a challenge for morphologically rich languages like Bangla. The complexity arises from the existence of a large number of alphabets, the presence of several diacritic forms, and the appearance of complex conjuncts. The difficulty is exacerbated by the fact that some graphemes occur infrequently but remain indispensable, so addressing the class imbalance is required for satisfactory results. This paper addresses this issue by introducing two knowledge distillation methods: Leveraging Isolated Letter Accumulations By Ordering Teacher Insights (LILA-BOTI) and Super Teacher LILA-BOTI. In both cases, a Convolutional Recurrent Neural Network (CRNN) student model is trained with the dark knowledge gained from a printed isolated character recognition teacher model. We conducted inter-dataset testing on \emph{BN-HTRd} and \emph{BanglaWriting} as our evaluation protocol, thus setting up a challenging problem where the results would better reflect the performance on unseen data. Our evaluations achieved up to a 3.5% increase in the F1-Macro score for the minor classes and up to 4.5% increase in our overall word recognition rate when compared with the base model (No KD) and conventional KD.

9.Enhanced Prototypical Learning for Unsupervised Domain Adaptation in LiDAR Semantic Segmentation ⬇️

Despite its importance, unsupervised domain adaptation (UDA) on LiDAR semantic segmentation is a task that has not received much attention from the research community. Only recently, a completion-based 3D method has been proposed to tackle the problem and formally set up the adaptive scenarios. However, the proposed pipeline is complex, voxel-based and requires multi-stage inference, which inhibits it for real-time inference. We propose a range image-based, effective and efficient method for solving UDA on LiDAR segmentation. The method exploits class prototypes from the source domain to pseudo label target domain pixels, which is a research direction showing good performance in UDA for natural image semantic segmentation. Applying such approaches to LiDAR scans has not been considered because of the severe domain shift and lack of pre-trained feature extractor that is unavailable in the LiDAR segmentation setup. However, we show that proper strategies, including reconstruction-based pre-training, enhanced prototypes, and selective pseudo labeling based on distance to prototypes, is sufficient enough to enable the use of prototypical approaches. We evaluate the performance of our method on the recently proposed LiDAR segmentation UDA scenarios. Our method achieves remarkable performance among contemporary methods.

10.Fine-Grained Counting with Crowd-Sourced Supervision ⬇️

Crowd-sourcing is an increasingly popular tool for image analysis in animal ecology. Computer vision methods that can utilize crowd-sourced annotations can help scale up analysis further. In this work we study the potential to do so on the challenging task of fine-grained counting. As opposed to the standard crowd counting task, fine-grained counting also involves classifying attributes of individuals in dense crowds. We introduce a new dataset from animal ecology to enable this study that contains 1.7M crowd-sourced annotations of 8 fine-grained classes. It is the largest available dataset for fine-grained counting and the first to enable the study of the task with crowd-sourced annotations. We introduce methods for generating aggregate "ground truths" from the collected annotations, as well as a counting method that can utilize the aggregate information. Our method improves results by 8% over a comparable baseline, indicating the potential for algorithms to learn fine-grained counting using crowd-sourced supervision.

11.Super Vision Transformer ⬇️

We attempt to reduce the computational costs in vision transformers (ViTs), which increase quadratically in the token number. We present a novel training paradigm that trains only one ViT model at a time, but is capable of providing improved image recognition performance with various computational costs. Here, the trained ViT model, termed super vision transformer (SuperViT), is empowered with the versatile ability to solve incoming patches of multiple sizes as well as preserve informative tokens with multiple keeping rates (the ratio of keeping tokens) to achieve good hardware efficiency for inference, given that the available hardware resources often change from time to time. Experimental results on ImageNet demonstrate that our SuperViT can considerably reduce the computational costs of ViT models with even performance increase. For example, we reduce 2x FLOPs of DeiT-S while increasing the Top-1 accuracy by 0.2% and 0.7% for 1.5x reduction. Also, our SuperViT significantly outperforms existing studies on efficient vision transformers. For example, when consuming the same amount of FLOPs, our SuperViT surpasses the recent state-of-the-art (SoTA) EViT by 1.1% when using DeiT-S as their backbones. The project of this work is made publicly available at this https URL.

12.Multi-Temporal Spatial-Spectral Comparison Network for Hyperspectral Anomalous Change Detection ⬇️

Hyperspectral anomalous change detection has been a challenging task for its emphasis on the dynamics of small and rare objects against the prevalent changes. In this paper, we have proposed a Multi-Temporal spatial-spectral Comparison Network for hyperspectral anomalous change detection (MTC-NET). The whole model is a deep siamese network, aiming at learning the prevalent spectral difference resulting from the complex imaging conditions from the hyperspectral images by contrastive learning. A three-dimensional spatial spectral attention module is designed to effectively extract the spatial semantic information and the key spectral differences. Then the gaps between the multi-temporal features are minimized, boosting the alignment of the semantic and spectral features and the suppression of the multi-temporal background spectral difference. The experiments on the "Viareggio 2013" datasets demonstrate the effectiveness of proposed MTC-NET.

13.Detection of Fights in Videos: A Comparison Study of Anomaly Detection and Action Recognition ⬇️

Detection of fights is an important surveillance application in videos. Most existing methods use supervised binary action recognition. Since frame-level annotations are very hard to get for anomaly detection, weakly supervised learning using multiple instance learning is widely used. This paper explores the detection of fights in videos as one special type of anomaly detection and as binary action recognition. We use the UBI-Fight and NTU-CCTV-Fight datasets for most of the study since they have frame-level annotations. We find that the anomaly detection has similar or even better performance than the action recognition. Furthermore, we study to use anomaly detection as a toolbox to generate training datasets for action recognition in an iterative way conditioned on the performance of the anomaly detection. Experiment results should show that we achieve state-of-the-art performance on three fight detection datasets.

14.Markedness in Visual Semantic AI ⬇️

We evaluate the state-of-the-art multimodal "visual semantic" model CLIP ("Contrastive Language Image Pretraining") for biases related to the marking of age, gender, and race or ethnicity. Given the option to label an image as "a photo of a person" or to select a label denoting race or ethnicity, CLIP chooses the "person" label 47.9% of the time for White individuals, compared with 5.0% or less for individuals who are Black, East Asian, Southeast Asian, Indian, or Latino or Hispanic. The model is more likely to rank the unmarked "person" label higher than labels denoting gender for Male individuals (26.7% of the time) vs. Female individuals (15.2% of the time). Age affects whether an individual is marked by the model: Female individuals under the age of 20 are more likely than Male individuals to be marked with a gender label, but less likely to be marked with an age label, while Female individuals over the age of 40 are more likely to be marked based on age than Male individuals. We also examine the self-similarity (mean pairwise cosine similarity) for each social group, where higher self-similarity denotes greater attention directed by CLIP to the shared characteristics (age, race, or gender) of the social group. As age increases, the self-similarity of representations of Female individuals increases at a higher rate than for Male individuals, with the disparity most pronounced at the "more than 70" age range. All ten of the most self-similar social groups are individuals under the age of 10 or over the age of 70, and six of the ten are Female individuals. Existing biases of self-similarity and markedness between Male and Female gender groups are further exacerbated when the groups compared are individuals who are White and Male and individuals who are Black and Female. Results indicate that CLIP reflects the biases of the language and society which produced its training data.

15.Towards Deeper Understanding of Camouflaged Object Detection ⬇️

Preys in the wild evolve to be camouflaged to avoid being recognized by predators. In this way, camouflage acts as a key defence mechanism across species that is critical to survival. To detect and segment the whole scope of a camouflaged object, camouflaged object detection (COD) is introduced as a binary segmentation task, with the binary ground truth camouflage map indicating the exact regions of the camouflaged objects. In this paper, we revisit this task and argue that the binary segmentation setting fails to fully understand the concept of camouflage. We find that explicitly modeling the conspicuousness of camouflaged objects against their particular backgrounds can not only lead to a better understanding about camouflage, but also provide guidance to designing more sophisticated camouflage techniques. Furthermore, we observe that it is some specific parts of camouflaged objects that make them detectable by predators. With the above understanding about camouflaged objects, we present the first triple-task learning framework to simultaneously localize, segment and rank camouflaged objects, indicating the conspicuousness level of camouflage. As no corresponding datasets exist for either the localization model or the ranking model, we generate localization maps with an eye tracker, which are then processed according to the instance level labels to generate our ranking-based training and testing dataset. We also contribute the largest COD testing set to comprehensively analyse performance of the camouflaged object detection models. Experimental results show that our triple-task learning framework achieves new state-of-the-art, leading to a more explainable camouflaged object detection network. Our code, data and results are available at: this https URL.

16.Towards automatic detection of wildlife trade using machine vision models ⬇️

Unsustainable trade in wildlife is one of the major threats affecting the global biodiversity crisis. An important part of the trade now occurs on the internet, especially on digital marketplaces and social media. Automated methods to identify trade posts are needed as resources for conservation are limited. Here, we developed machine vision models based on Deep Neural Networks with the aim to automatically identify images of exotic pet animals for sale. A new training dataset representing exotic pet animals advertised for sale on the web was generated for this purpose. We trained 24 neural-net models spanning a combination of five different architectures, three methods of training and two types of datasets. Specifically, model generalisation improved after setting a portion of the training images to represent negative features. Models were evaluated on both within and out of distribution data to test wider model applicability. The top performing models achieved an f-score of over 0.95 on within distribution evaluation and between 0.75 to 0.87 on the two out of distribution datasets. Notably, feature visualisation indicated that models performed well in detecting the surrounding context (e.g. a cage) in which an animal was located, therefore helping to automatically detect images of animals in non-natural environments. The proposed methods can help investigate the online wildlife trade, but can also be adapted to study other types of people-nature interactions from digital platforms. Future studies can use these findings to build robust machine learning models and new data collection pipelines for more taxonomic groups.

17.Continual Barlow Twins: continual self-supervised learning for remote sensing semantic segmentation ⬇️

In the field of Earth Observation (EO), Continual Learning (CL) algorithms have been proposed to deal with large datasets by decomposing them into several subsets and processing them incrementally. The majority of these algorithms assume that data is (a) coming from a single source, and (b) fully labeled. Real-world EO datasets are instead characterized by a large heterogeneity (e.g., coming from aerial, satellite, or drone scenarios), and for the most part they are unlabeled, meaning they can be fully exploited only through the emerging Self-Supervised Learning (SSL) paradigm. For these reasons, in this paper we propose a new algorithm for merging SSL and CL for remote sensing applications, that we call Continual Barlow Twins (CBT). It combines the advantages of one of the simplest self-supervision techniques, i.e., Barlow Twins, with the Elastic Weight Consolidation method to avoid catastrophic forgetting. In addition, for the first time we evaluate SSL methods on a highly heterogeneous EO dataset, showing the effectiveness of these strategies on a novel combination of three almost non-overlapping domains datasets (airborne Potsdam dataset, satellite US3D dataset, and drone UAVid dataset), on a crucial downstream task in EO, i.e., semantic segmentation. Encouraging results show the superiority of SSL in this setting, and the effectiveness of creating an incremental effective pretrained feature extractor, based on ResNet50, without the need of relying on the complete availability of all the data, with a valuable saving of time and resources.

18.SelfReformer: Self-Refined Network with Transformer for Salient Object Detection ⬇️

The global and local contexts significantly contribute to the integrity of predictions in Salient Object Detection (SOD). Unfortunately, existing methods still struggle to generate complete predictions with fine details. There are two major problems in conventional approaches: first, for global context, high-level CNN-based encoder features cannot effectively catch long-range dependencies, resulting in incomplete predictions. Second, downsampling the ground truth to fit the size of predictions will introduce inaccuracy as the ground truth details are lost during interpolation or pooling. Thus, in this work, we developed a Transformer-based network and framed a supervised task for a branch to learn the global context information explicitly. Besides, we adopt Pixel Shuffle from Super-Resolution (SR) to reshape the predictions back to the size of ground truth instead of the reverse. Thus details in the ground truth are untouched. In addition, we developed a two-stage Context Refinement Module (CRM) to fuse global context and automatically locate and refine the local details in the predictions. The proposed network can guide and correct itself based on the global and local context generated, thus is named, Self-Refined Transformer (SelfReformer). Extensive experiments and evaluation results on five benchmark datasets demonstrate the outstanding performance of the network, and we achieved the state-of-the-art.

19.GR-GAN: Gradual Refinement Text-to-image Generation ⬇️

A good Text-to-Image model should not only generate high quality images, but also ensure the consistency between the text and the generated image. Previous models failed to simultaneously fix both sides well. This paper proposes a Gradual Refinement Generative Adversarial Network (GR-GAN) to alleviates the problem efficiently. A GRG module is designed to generate images from low resolution to high resolution with the corresponding text constraints from coarse granularity (sentence) to fine granularity (word) stage by stage, a ITM module is designed to provide image-text matching losses at both sentence-image level and word-region level for corresponding stages. We also introduce a new metric Cross-Model Distance (CMD) for simultaneously evaluating image quality and image-text consistency. Experimental results show GR-GAN significant outperform previous models, and achieve new state-of-the-art on both FID and CMD. A detailed analysis demonstrates the efficiency of different generation stages in GR-GAN.

20.Dynamic Split Computing for Efficient Deep Edge Intelligence ⬇️

Deploying deep neural networks (DNNs) on IoT and mobile devices is a challenging task due to their limited computational resources. Thus, demanding tasks are often entirely offloaded to edge servers which can accelerate inference, however, it also causes communication cost and evokes privacy concerns. In addition, this approach leaves the computational capacity of end devices unused. Split computing is a paradigm where a DNN is split into two sections; the first section is executed on the end device, and the output is transmitted to the edge server where the final section is executed. Here, we introduce dynamic split computing, where the optimal split location is dynamically selected based on the state of the communication channel. By using natural bottlenecks that already exist in modern DNN architectures, dynamic split computing avoids retraining and hyperparameter optimization, and does not have any negative impact on the final accuracy of DNNs. Through extensive experiments, we show that dynamic split computing achieves faster inference in edge computing environments where the data rate and server load vary over time.

21.What You See is What You Classify: Black Box Attributions ⬇️

An important step towards explaining deep image classifiers lies in the identification of image regions that contribute to individual class scores in the model's output. However, doing this accurately is a difficult task due to the black-box nature of such networks. Most existing approaches find such attributions either using activations and gradients or by repeatedly perturbing the input. We instead address this challenge by training a second deep network, the Explainer, to predict attributions for a pre-trained black-box classifier, the Explanandum. These attributions are in the form of masks that only show the classifier-relevant parts of an image, masking out the rest. Our approach produces sharper and more boundary-precise masks when compared to the saliency maps generated by other methods. Moreover, unlike most existing approaches, ours is capable of directly generating very distinct class-specific masks. Finally, the proposed method is very efficient for inference since it only takes a single forward pass through the Explainer to generate all class-specific masks. We show that our attributions are superior to established methods both visually and quantitatively, by evaluating them on the PASCAL VOC-2007 and Microsoft COCO-2014 datasets.

22.[Re] Distilling Knowledge via Knowledge Review ⬇️

This effort aims to reproduce the results of experiments and analyze the robustness of the review framework for knowledge distillation introduced in the CVPR '21 paper 'Distilling Knowledge via Knowledge Review' by Chen et al. Previous works in knowledge distillation only studied connections paths between the same levels of the student and the teacher, and cross-level connection paths had not been considered. Chen et al. propose a new residual learning framework to train a single student layer using multiple teacher layers. They also design a novel fusion module to condense feature maps across levels and a loss function to compare feature information stored across different levels to improve performance. In this work, we consistently verify the improvements in test accuracy across student models as reported in the original paper and study the effectiveness of the novel modules introduced by conducting ablation studies and new experiments.

23.Fusing Multiscale Texture and Residual Descriptors for Multilevel 2D Barcode Rebroadcasting Detection ⬇️

Nowadays, 2D barcodes have been widely used for advertisement, mobile payment, and product authentication. However, in applications related to product authentication, an authentic 2D barcode can be illegally copied and attached to a counterfeited product in such a way to bypass the authentication scheme. In this paper, we employ a proprietary 2D barcode pattern and use multimedia forensics methods to analyse the scanning and printing artefacts resulting from the copy (rebroadcasting) attack. A diverse and complementary feature set is proposed to quantify the barcode texture distortions introduced during the illegal copying process. The proposed features are composed of global and local descriptors, which characterize the multi-scale texture appearance and the points of interest distribution, respectively. The proposed descriptors are compared against some existing texture descriptors and deep learning-based approaches under various scenarios, such as cross-datasets and cross-size. Experimental results highlight the practicality of the proposed method in real-world settings.

24.A Novel Face-Anti Spoofing Neural Network Model For Face Recognition And Detection ⬇️

Face Recognition (FR) systems are being used in a variety of applications, including road crossings, banking, and mobile banking. The widespread use of FR systems has raised concerns about the safety of face biometrics against spoofing attacks, which use the use of a photo or video of a legitimate user's face to gain illegal access to the resources or activities. Despite the development of several FAS or liveness detection methods (which determine whether a face is live or spoofed at the time of acquisition), the problem remains unsolved due to the difficulty of identifying discrimination and operationally reasonably priced spoof characteristics but also approaches. Additionally, certain facial portions are frequently repeated or correlate to image clutter, resulting in poor performance overall. This research proposes a face-anti-spoofing neural network model that outperforms existing models and has an efficiency of 0.89 percent.

25.Vision Transformer: Vit and its Derivatives ⬇️

Transformer, an attention-based encoder-decoder architecture, has not only revolutionized the field of natural language processing (NLP), but has also done some pioneering work in the field of computer vision (CV). Compared to convolutional neural networks (CNNs), the Vision Transformer (ViT) relies on excellent modeling capabilities to achieve very good performance on several benchmarks such as ImageNet, COCO, and ADE20k. ViT is inspired by the self-attention mechanism in natural language processing, where word embeddings are replaced with patch embeddings.
This paper reviews the derivatives of ViT and the cross-applications of ViT with other fields.

26.Hyperspectral Image Classification With Contrastive Graph Convolutional Network ⬇️

Recently, Graph Convolutional Network (GCN) has been widely used in Hyperspectral Image (HSI) classification due to its satisfactory performance. However, the number of labeled pixels is very limited in HSI, and thus the available supervision information is usually insufficient, which will inevitably degrade the representation ability of most existing GCN-based methods. To enhance the feature representation ability, in this paper, a GCN model with contrastive learning is proposed to explore the supervision signals contained in both spectral information and spatial relations, which is termed Contrastive Graph Convolutional Network (ConGCN), for HSI classification. First, in order to mine sufficient supervision signals from spectral information, a semi-supervised contrastive loss function is utilized to maximize the agreement between different views of the same node or the nodes from the same land cover category. Second, to extract the precious yet implicit spatial relations in HSI, a graph generative loss function is leveraged to explore supplementary supervision signals contained in the graph topology. In addition, an adaptive graph augmentation technique is designed to flexibly incorporate the spectral-spatial priors of HSI, which helps facilitate the subsequent contrastive representation learning. The extensive experimental results on four typical benchmark datasets firmly demonstrate the effectiveness of the proposed ConGCN in both qualitative and quantitative aspects.

27.2-d signature of images and texture classification ⬇️

We introduce a proper notion of 2-dimensional signature for images. This object is inspired by the so-called rough paths theory, and it captures many essential features of a 2-dimensional object such as an image. It thus serves as a low-dimensional feature for pattern classification. Here we implement a simple procedure for texture classification. In this context, we show that a low dimensional set of features based on signatures produces an excellent accuracy.

28.Deep Neural Network approaches for Analysing Videos of Music Performances ⬇️

This paper presents a framework to automate the labelling process for gestures in musical performance videos with a 3D Convolutional Neural Network (CNN). While this idea was proposed in a previous study, this paper introduces several novelties: (i) Presents a novel method to overcome the class imbalance challenge and make learning possible for co-existent gestures by batch balancing approach and spatial-temporal representations of gestures. (ii) Performs a detailed study on 7 and 18 categories of gestures generated during the performance (guitar play) of musical pieces that have been video-recorded. (iii) Investigates the possibility to use audio features. (iv) Extends the analysis to multiple videos. The novel methods significantly improve the performance of gesture identification by 12 %, when compared to the previous work (51 % in this study over 39 % in previous work). We successfully validate the proposed methods on 7 super classes (72 %), an ensemble of the 18 gestures/classes, and additional videos (75 %).

29.Geocentric Pose Analysis of Satellite Imagery Using Deep Learning ⬇️

Roughly 6,800 natural disasters occur worldwide annually, and this alarming number continues to grow due to the effects of climate change. Effective methods to improve natural disaster response include performing change detection, map alignment, and vision-aided navigation to allow for the time-efficient delivery of life-saving aid. Current software functions optimally only on nadir images taken ninety degrees above ground level. The inability to generalize to oblique images increases the need to compute an image's geocentric pose, which is its spatial orientation with respect to gravity. This Deep Learning investigation presents three convolutional models to predict geocentric pose using 5,923 nadir and oblique RGB satellite images of cities worldwide. The first model is an autoencoder that condenses the 256 x 256 x 3 images to 32 x 32 x 16 latent space representations, demonstrating the ability to learn useful features from the data. The second model is a U-Net Fully Convolutional Network with skip connections used to predict each image's corresponding pixel-level elevation mask. This model achieves a median absolute deviation of 0.335 meters and an R2 of 0.865 on test data. Afterward, the elevation masks are concatenated with the RGB images to form 256 x 256 x 4 inputs of the third model, which predicts each image's rotation angle and scale, the components of its geocentric pose. This Deep Convolutional Neural Network achieves an R2 of 0.904 on test data, significantly outperforming previous models designed by researchers. The high-accuracy software built in this study contributes to crucial procedures that can accelerate disaster relief and save human lives.

30.Scalable Kernel-Based Minimum Mean Square Error Estimator for Accelerated Image Error Concealment ⬇️

Error concealment is of great importance for block-based video systems, such as DVB or video streaming services. In this paper, we propose a novel scalable spatial error concealment algorithm that aims at obtaining high quality reconstructions with reduced computational burden. The proposed technique exploits the excellent reconstructing abilities of the kernel-based minimum mean square error K-MMSE estimator. We propose to decompose this approach into a set of hierarchically stacked layers. The first layer performs the basic reconstruction that the subsequent layers can eventually refine. In addition, we design a layer management mechanism, based on profiles, that dynamically adapts the use of higher layers to the visual complexity of the area being reconstructed. The proposed technique outperforms other state-of-the-art algorithms and produces high quality reconstructions, equivalent to K-MMSE, while requiring around one tenth of its computational time.

31.Around View Monitoring System for Hydraulic Excavators ⬇️

This paper describes the Around View Monitoring (AVM) system for hydraulic excavators that prevents the safety accidents caused by blind spots and increases the operational efficiency. To verify the developed system, experiments were conducted with its prototype. The experimental results demonstrate its applicability in the field with the following values: 7m of a visual range, 15fps of image refresh rate, 300ms of working information data reception rate, and 300ms of surface condition data reception rate.

32.Denoising-based image reconstruction from pixels located at non-integer positions ⬇️

Digital images are commonly represented as regular 2D arrays, so pixels are organized in form of a matrix addressed by integers. However, there are many image processing operations, such as rotation or motion compensation, that produce pixels at non-integer positions. Typically, image reconstruction techniques cannot handle samples at non-integer positions. In this paper, we propose to use triangulation-based reconstruction as initial estimate that is later refined by a novel adaptive denoising framework. Simulations reveal that improvements of up to more than 1.8 dB (in terms of PSNR) are achieved with respect to the initial estimate.

33.Feature-Distribution Perturbation and Calibration for Generalized Person ReID ⬇️

Person Re-identification (ReID) has been advanced remarkably over the last 10 years along with the rapid development of deep learning for visual recognition. However, the i.i.d. (independent and identically distributed) assumption commonly held in most deep learning models is somewhat non-applicable to ReID considering its objective to identify images of the same pedestrian across cameras at different locations often of variable and independent domain characteristics that are also subject to view-biased data distribution. In this work, we propose a Feature-Distribution Perturbation and Calibration (PECA) method to derive generic feature representations for person ReID, which is not only discriminative across cameras but also agnostic and deployable to arbitrary unseen target domains. Specifically, we perform per-domain feature-distribution perturbation to refrain the model from overfitting to the domain-biased distribution of each source (seen) domain by enforcing feature invariance to distribution shifts caused by perturbation. Furthermore, we design a global calibration mechanism to align feature distributions across all the source domains to improve the model generalization capacity by eliminating domain bias. These local perturbation and global calibration are conducted simultaneously, which share the same principle to avoid models overfitting by regularization respectively on the perturbed and the original distributions. Extensive experiments were conducted on eight person ReID datasets and the proposed PECA model outperformed the state-of-the-art competitors by significant margins.

34.Deep Image Retrieval is not Robust to Label Noise ⬇️

Large-scale datasets are essential for the success of deep learning in image retrieval. However, manual assessment errors and semi-supervised annotation techniques can lead to label noise even in popular datasets. As previous works primarily studied annotation quality in image classification tasks, it is still unclear how label noise affects deep learning approaches to image retrieval. In this work, we show that image retrieval methods are less robust to label noise than image classification ones. Furthermore, we, for the first time, investigate different types of label noise specific to image retrieval tasks and study their effect on model performance.

35.Active Domain Adaptation with Multi-level Contrastive Units for Semantic Segmentation ⬇️

To further reduce the cost of semi-supervised domain adaptation (SSDA) labeling, a more effective way is to use active learning (AL) to annotate a selected subset with specific properties. However, domain adaptation tasks are always addressed in two interactive aspects: domain transfer and the enhancement of discrimination, which requires the selected data to be both uncertain under the model and diverse in feature space. Contrary to active learning in classification tasks, it is usually challenging to select pixels that contain both the above properties in segmentation tasks, leading to the complex design of pixel selection strategy. To address such an issue, we propose a novel Active Domain Adaptation scheme with Multi-level Contrastive Units (ADA-MCU) for semantic image segmentation. A simple pixel selection strategy followed with the construction of multi-level contrastive units is introduced to optimize the model for both domain adaptation and active supervised learning. In practice, MCUs are constructed from intra-image, cross-image, and cross-domain levels by using both labeled and unlabeled pixels. At each level, we define contrastive losses from center-to-center and pixel-to-pixel manners, with the aim of jointly aligning the category centers and reducing outliers near the decision boundaries. In addition, we also introduce a categories correlation matrix to implicitly describe the relationship between categories, which are used to adjust the weights of the losses for MCUs. Extensive experimental results on standard benchmarks show that the proposed method achieves competitive performance against state-of-the-art SSDA methods with 50% fewer labeled pixels and significantly outperforms state-of-the-art with a large margin by using the same level of annotation cost.

36.NPU-BOLT: A Dataset for Bolt Object Detection in Natural Scene Images ⬇️

Bolt joints are very common and important in engineering structures. Due to extreme service environment and load factors, bolts often get loose or even disengaged. To real-time or timely detect the loosed or disengaged bolts is an urgent need in practical engineering, which is critical to keep structural safety and service life. In recent years, many bolt loosening detection methods using deep learning and machine learning techniques have been proposed and are attracting more and more attention. However, most of these studies use bolt images captured in laboratory for deep leaning model training. The images are obtained in a well-controlled light, distance, and view angle conditions. Also, the bolted structures are well designed experimental structures with brand new bolts and the bolts are exposed without any shelter nearby. It is noted that in practical engineering, the above well controlled lab conditions are not easy realized and the real bolt images often have blur edges, oblique perspective, partial occlusion and indistinguishable colors etc., which make the trained models obtained in laboratory conditions loss their accuracy or fails. Therefore, the aim of this study is to develop a dataset named NPU-BOLT for bolt object detection in natural scene images and open it to researchers for public use and further development. In the first version of the dataset, it contains 337 samples of bolt joints images mainly in the natural environment, with image data sizes ranging from 400400 to 60004000, totaling approximately 1275 bolt targets. The bolt targets are annotated into four categories named blur bolt, bolt head, bolt nut and bolt side. The dataset is tested with advanced object detection models including yolov5, Faster-RCNN and CenterNet. The effectiveness of the dataset is validated.

37.Online Hybrid Lightweight Representations Learning: Its Application to Visual Tracking ⬇️

This paper presents a novel hybrid representation learning framework for streaming data, where an image frame in a video is modeled by an ensemble of two distinct deep neural networks; one is a low-bit quantized network and the other is a lightweight full-precision network. The former learns coarse primary information with low cost while the latter conveys residual information for high fidelity to original representations. The proposed parallel architecture is effective to maintain complementary information since fixed-point arithmetic can be utilized in the quantized network and the lightweight model provides precise representations given by a compact channel-pruned network. We incorporate the hybrid representation technique into an online visual tracking task, where deep neural networks need to handle temporal variations of target appearances in real-time. Compared to the state-of-the-art real-time trackers based on conventional deep neural networks, our tracking algorithm demonstrates competitive accuracy on the standard benchmarks with a small fraction of computational cost and memory footprint.

38.PEVL: Position-enhanced Pre-training and Prompt Tuning for Vision-language Models ⬇️

Vision-language pre-training (VLP) has shown impressive performance on a wide range of cross-modal tasks, where VLP models without reliance on object detectors are becoming the mainstream due to their superior computation efficiency and competitive performance. However, the removal of object detectors also deprives the capability of VLP models in explicit object modeling, which is essential to various position-sensitive vision-language (VL) tasks, such as referring expression comprehension and visual commonsense reasoning. To address the challenge, we introduce PEVL that enhances the pre-training and prompt tuning of VLP models with explicit object position modeling. Specifically, PEVL reformulates discretized object positions and language in a unified language modeling framework, which facilitates explicit VL alignment during pre-training, and also enables flexible prompt tuning for various downstream tasks. We show that PEVL enables state-of-the-art performance of detector-free VLP models on position-sensitive tasks such as referring expression comprehension and phrase grounding, and also improves the performance on position-insensitive tasks with grounded inputs. We make the data and code for this paper publicly available at this https URL.

39.OPQ: Compressing Deep Neural Networks with One-shot Pruning-Quantization ⬇️

As Deep Neural Networks (DNNs) usually are overparameterized and have millions of weight parameters, it is challenging to deploy these large DNN models on resource-constrained hardware platforms, e.g., smartphones. Numerous network compression methods such as pruning and quantization are proposed to reduce the model size significantly, of which the key is to find suitable compression allocation (e.g., pruning sparsity and quantization codebook) of each layer. Existing solutions obtain the compression allocation in an iterative/manual fashion while finetuning the compressed model, thus suffering from the efficiency issue. Different from the prior art, we propose a novel One-shot Pruning-Quantization (OPQ) in this paper, which analytically solves the compression allocation with pre-trained weight parameters only. During finetuning, the compression module is fixed and only weight parameters are updated. To our knowledge, OPQ is the first work that reveals pre-trained model is sufficient for solving pruning and quantization simultaneously, without any complex iterative/manual optimization at the finetuning stage. Furthermore, we propose a unified channel-wise quantization method that enforces all channels of each layer to share a common codebook, which leads to low bit-rate allocation without introducing extra overhead brought by traditional channel-wise quantization. Comprehensive experiments on ImageNet with AlexNet/MobileNet-V1/ResNet-50 show that our method improves accuracy and training efficiency while obtains significantly higher compression rates compared to the state-of-the-art.

40.Heterogeneous Semantic Transfer for Multi-label Recognition with Partial Labels ⬇️

Multi-label image recognition with partial labels (MLR-PL), in which some labels are known while others are unknown for each image, may greatly reduce the cost of annotation and thus facilitate large-scale MLR. We find that strong semantic correlations exist within each image and across different images, and these correlations can help transfer the knowledge possessed by the known labels to retrieve the unknown labels and thus improve the performance of the MLR-PL task (see Figure 1). In this work, we propose a novel heterogeneous semantic transfer (HST) framework that consists of two complementary transfer modules that explore both within-image and cross-image semantic correlations to transfer the knowledge possessed by known labels to generate pseudo labels for the unknown labels. Specifically, an intra-image semantic transfer (IST) module learns an image-specific label co-occurrence matrix for each image and maps the known labels to complement the unknown labels based on these matrices. Additionally, a cross-image transfer (CST) module learns category-specific feature-prototype similarities and then helps complement the unknown labels that have high degrees of similarity with the corresponding prototypes. Finally, both the known and generated pseudo labels are used to train MLR models. Extensive experiments conducted on the Microsoft COCO, Visual Genome, and Pascal VOC 2007 datasets show that the proposed HST framework achieves superior performance to that of current state-of-the-art algorithms. Specifically, it obtains mean average precision (mAP) improvements of 1.4%, 3.3%, and 0.4% on the three datasets over the results of the best-performing previously developed algorithm.

41.ConvPoseCNN2: Prediction and Refinement of Dense 6D Object Poses ⬇️

Object pose estimation is a key perceptual capability in robotics. We propose a fully-convolutional extension of the PoseCNN method, which densely predicts object translations and orientations. This has several advantages such as improving the spatial resolution of the orientation predictions -- useful in highly-cluttered arrangements, significant reduction in parameters by avoiding full connectivity, and fast inference. We propose and discuss several aggregation methods for dense orientation predictions that can be applied as a post-processing step, such as averaging and clustering techniques. We demonstrate that our method achieves the same accuracy as PoseCNN on the challenging YCB-Video dataset and provide a detailed ablation study of several variants of our method. Finally, we demonstrate that the model can be further improved by inserting an iterative refinement module into the middle of the network, which enforces consistency of the prediction.

42.Gradient Hedging for Intensively Exploring Salient Interpretation beyond Neuron Activation ⬇️

Hedging is a strategy for reducing the potential risks in various types of investments by adopting an opposite position in a related asset. Motivated by the equity technique, we introduce a method for decomposing output predictions into intensive salient attributions by hedging the evidence for a decision. We analyze the conventional approach applied to the evidence for a decision and discuss the paradox of the conservation rule. Subsequently, we define the viewpoint of evidence as a gap of positive and negative influence among the gradient-derived initial contribution maps and propagate the antagonistic elements to the evidence as suppressors, following the criterion of the degree of positive attribution defined by user preference. In addition, we reflect the severance or sparseness contribution of inactivated neurons, which are mostly irrelevant to a decision, resulting in increased robustness to interpretability. We conduct the following assessments in a verified experimental environment: pointing game, most relevant first region insertion, outside-inside relevance ratio, and mean average precision on the PASCAL VOC 2007, MS COCO 2014, and ImageNet datasets. The results demonstrate that our method outperforms existing attribution methods in distinctive, intensive, and intuitive visualization with robustness and applicability in general models.

43.Paddy Doctor: A Visual Image Dataset for Paddy Disease Classification ⬇️

One of the critical biotic stress factors paddy farmers face is diseases caused by bacteria, fungi, and other organisms. These diseases affect plants' health severely and lead to significant crop loss. Most of these diseases can be identified by regularly observing the leaves and stems under expert supervision. In a country with vast agricultural regions and limited crop protection experts, manual identification of paddy diseases is challenging. Thus, to add a solution to this problem, it is necessary to automate the disease identification process and provide easily accessible decision support tools to enable effective crop protection measures. However, the lack of availability of public datasets with detailed disease information limits the practical implementation of accurate disease detection systems. This paper presents Paddy Doctor, a visual image dataset for identifying paddy diseases. Our dataset contains 13,876 annotated paddy leaf images across ten classes (nine diseases and normal leaf). We benchmarked the Paddy Doctor using a Convolutional Neural Network (CNN) and two transfer learning approaches, VGG16 and MobileNet. The experimental results show that MobileNet achieves the highest classification accuracy of 93.83%. We release our dataset and reproducible code in the open source for community use.

44.Supporting Vision-Language Model Inference with Causality-pruning Knowledge Prompt ⬇️

Vision-language models are pre-trained by aligning image-text pairs in a common space so that the models can deal with open-set visual concepts by learning semantic information from textual labels. To boost the transferability of these models on downstream tasks in a zero-shot manner, recent works explore generating fixed or learnable prompts, i.e., classification weights are synthesized from natural language describing task-relevant categories, to reduce the gap between tasks in the training and test phases. However, how and what prompts can improve inference performance remains unclear. In this paper, we explicitly provide exploration and clarify the importance of including semantic information in prompts, while existing prompt methods generate prompts without exploring the semantic information of textual labels. A challenging issue is that manually constructing prompts, with rich semantic information, requires domain expertise and is extremely time-consuming. To this end, we propose Causality-pruning Knowledge Prompt (CapKP) for adapting pre-trained vision-language models to downstream image recognition. CapKP retrieves an ontological knowledge graph by treating the textual label as a query to explore task-relevant semantic information. To further refine the derived semantic information, CapKP introduces causality-pruning by following the first principle of Granger causality. Empirically, we conduct extensive evaluations to demonstrate the effectiveness of CapKP, e.g., with 8 shots, CapKP outperforms the manual-prompt method by 12.51% and the learnable-prompt method by 1.39% on average, respectively. Experimental analyses prove the superiority of CapKP in domain generalization compared to benchmark approaches.

45.PointDistiller: Structured Knowledge Distillation Towards Efficient and Compact 3D Detection ⬇️

The remarkable breakthroughs in point cloud representation learning have boosted their usage in real-world applications such as self-driving cars and virtual reality. However, these applications usually have an urgent requirement for not only accurate but also efficient 3D object detection. Recently, knowledge distillation has been proposed as an effective model compression technique, which transfers the knowledge from an over-parameterized teacher to a lightweight student and achieves consistent effectiveness in 2D vision. However, due to point clouds' sparsity and irregularity, directly applying previous image-based knowledge distillation methods to point cloud detectors usually leads to unsatisfactory performance. To fill the gap, this paper proposes PointDistiller, a structured knowledge distillation framework for point clouds-based 3D detection. Concretely, PointDistiller includes local distillation which extracts and distills the local geometric structure of point clouds with dynamic graph convolution and reweighted learning strategy, which highlights student learning on the crucial points or voxels to improve knowledge distillation efficiency. Extensive experiments on both voxels-based and raw points-based detectors have demonstrated the effectiveness of our method over seven previous knowledge distillation methods. For instance, our 4X compressed PointPillars student achieves 2.8 and 3.4 mAP improvements on BEV and 3D object detection, outperforming its teacher by 0.9 and 1.8 mAP, respectively. Codes have been released at this https URL.

46.FaceMAE: Privacy-Preserving Face Recognition via Masked Autoencoders ⬇️

Face recognition, as one of the most successful applications in artificial intelligence, has been widely used in security, administration, advertising, and healthcare. However, the privacy issues of public face datasets have attracted increasing attention in recent years. Previous works simply mask most areas of faces or synthesize samples using generative models to construct privacy-preserving face datasets, which overlooks the trade-off between privacy protection and data utility. In this paper, we propose a novel framework FaceMAE, where the face privacy and recognition performance are considered simultaneously. Firstly, randomly masked face images are used to train the reconstruction module in FaceMAE. We tailor the instance relation matching (IRM) module to minimize the distribution gap between real faces and FaceMAE reconstructed ones. During the deployment phase, we use trained FaceMAE to reconstruct images from masked faces of unseen identities without extra training. The risk of privacy leakage is measured based on face retrieval between reconstructed and original datasets. Experiments prove that the identities of reconstructed images are difficult to be retrieved. We also perform sufficient privacy-preserving face recognition on several public face datasets (i.e. CASIA-WebFace and WebFace260M). Compared to previous state of the arts, FaceMAE consistently \textbf{reduces at least 50% error rate} on LFW, CFP-FP and AgeDB.

47.MonoFormer: Towards Generalization of self-supervised monocular depth estimation with Transformers ⬇️

Self-supervised monocular depth estimation has been widely studied recently. Most of the work has focused on improving performance on benchmark datasets, such as KITTI, but has offered a few experiments on generalization performance. In this paper, we investigate the backbone networks (e.g. CNNs, Transformers, and CNN-Transformer hybrid models) toward the generalization of monocular depth estimation. We first evaluate state-of-the-art models on diverse public datasets, which have never been seen during the network training. Next, we investigate the effects of texture-biased and shape-biased representations using the various texture-shifted datasets that we generated. We observe that Transformers exhibit a strong shape bias and CNNs do a strong texture-bias. We also find that shape-biased models show better generalization performance for monocular depth estimation compared to texture-biased models. Based on these observations, we newly design a CNN-Transformer hybrid network with a multi-level adaptive feature fusion module, called MonoFormer. The design intuition behind MonoFormer is to increase shape bias by employing Transformers while compensating for the weak locality bias of Transformers by adaptively fusing multi-level representations. Extensive experiments show that the proposed method achieves state-of-the-art performance with various public datasets. Our method also shows the best generalization ability among the competitive methods.

48.Self-distilled Knowledge Delegator for Exemplar-free Class Incremental Learning ⬇️

Exemplar-free incremental learning is extremely challenging due to inaccessibility of data from old tasks. In this paper, we attempt to exploit the knowledge encoded in a previously trained classification model to handle the catastrophic forgetting problem in continual learning. Specifically, we introduce a so-called knowledge delegator, which is capable of transferring knowledge from the trained model to a randomly re-initialized new model by generating informative samples. Given the previous model only, the delegator is effectively learned using a self-distillation mechanism in a data-free manner. The knowledge extracted by the delegator is then utilized to maintain the performance of the model on old tasks in incremental learning. This simple incremental learning framework surpasses existing exemplar-free methods by a large margin on four widely used class incremental benchmarks, namely CIFAR-100, ImageNet-Subset, Caltech-101 and Flowers-102. Notably, we achieve comparable performance to some exemplar-based methods without accessing any exemplars.

49.Saliency-Driven Active Contour Model for Image Segmentation ⬇️

Active contour models have achieved prominent success in the area of image segmentation, allowing complex objects to be segmented from the background for further analysis. Existing models can be divided into region-based active contour models and edge-based active contour models. However, both models use direct image data to achieve segmentation and face many challenging problems in terms of the initial contour position, noise sensitivity, local minima and inefficiency owing to the in-homogeneity of image intensities. The saliency map of an image changes the image representation, making it more visual and meaningful. In this study, we propose a novel model that uses the advantages of a saliency map with local image information (LIF) and overcomes the drawbacks of previous models. The proposed model is driven by a saliency map of an image and the local image information to enhance the progress of the active contour models. In this model, the saliency map of an image is first computed to find the saliency driven local fitting energy. Then, the saliency-driven local fitting energy is combined with the LIF model, resulting in a final novel energy functional. This final energy functional is formulated through a level set formulation, and regulation terms are added to evolve the contour more precisely across the object boundaries. The quality of the proposed method was verified on different synthetic images, real images and publicly available datasets, including medical images. The image segmentation results, and quantitative comparisons confirmed the contour initialization independence, noise insensitivity, and superior segmentation accuracy of the proposed model in comparison to the other segmentation models.

50.Vegetation Mapping by UAV Visible Imagery and Machine Learning ⬇️

An experimental field cropped with sugar-beet with a wide spreading of weeds has been used to test vegetation identification from drone visible imagery. Expert masked and hue-filtered pictures have been used to train several Machine Learning algorithms to develop a semi-automatic methodology for identification and mapping species at high resolution. Results show that 5m altitude allows for obtaining maps with an identification efficiency of more than 90%. Such a method can be easily integrated to present VRHA, as much as tools to obtain detailed maps of vegetation.

51.Keypoint-Based Category-Level Object Pose Tracking from an RGB Sequence with Uncertainty Estimation ⬇️

We propose a single-stage, category-level 6-DoF pose estimation algorithm that simultaneously detects and tracks instances of objects within a known category. Our method takes as input the previous and current frame from a monocular RGB video, as well as predictions from the previous frame, to predict the bounding cuboid and 6-DoF pose (up to scale). Internally, a deep network predicts distributions over object keypoints (vertices of the bounding cuboid) in image coordinates, after which a novel probabilistic filtering process integrates across estimates before computing the final pose using PnP. Our framework allows the system to take previous uncertainties into consideration when predicting the current frame, resulting in predictions that are more accurate and stable than single frame methods. Extensive experiments show that our method outperforms existing approaches on the challenging Objectron benchmark of annotated object videos. We also demonstrate the usability of our work in an augmented reality setting.

52.Body Composition Estimation Based on Multimodal Multi-task Deep Neural Network ⬇️

In addition to body weight and Body Mass Index (BMI), body composition is an essential data point that allows people to understand their overall health and body fitness. However, body composition is largely made up of muscle, fat, bones, and water, which makes estimation not as easy and straightforward as measuring body weight. In this paper, we introduce a multimodal multi-task deep neural network to estimate body fat percentage and skeletal muscle mass by analyzing facial images in addition to a person's height, gender, age, and weight information. Using a dataset representative of demographics in Japan, we confirmed that the proposed approach performed better compared to the existing methods. Moreover, the multi-task approach implemented in this study is also able to grasp the negative correlation between body fat percentage and skeletal muscle mass gain/loss.

53.RCP: Recurrent Closest Point for Scene Flow Estimation on 3D Point Clouds ⬇️

3D motion estimation including scene flow and point cloud registration has drawn increasing interest. Inspired by 2D flow estimation, recent methods employ deep neural networks to construct the cost volume for estimating accurate 3D flow. However, these methods are limited by the fact that it is difficult to define a search window on point clouds because of the irregular data structure. In this paper, we avoid this irregularity by a simple yet effective method.We decompose the problem into two interlaced stages, where the 3D flows are optimized point-wisely at the first stage and then globally regularized in a recurrent network at the second stage. Therefore, the recurrent network only receives the regular point-wise information as the input. In the experiments, we evaluate the proposed method on both the 3D scene flow estimation and the point cloud registration task. For 3D scene flow estimation, we make comparisons on the widely used FlyingThings3D and KITTIdatasets. For point cloud registration, we follow previous works and evaluate the data pairs with large pose and partially overlapping from ModelNet40. The results show that our method outperforms the previous method and achieves a new state-of-the-art performance on both 3D scene flow estimation and point cloud registration, which demonstrates the superiority of the proposed zero-order method on irregular point cloud data.

54.LexiconNet: An End-to-End Handwritten Paragraph Text Recognition System ⬇️

Historical documents present in the form of libraries needs to be digitised. The recognition of these unconstrained cursive handwritten documents is a challenging task. In the present work, neural network based classifier is used. The recognition of scanned document images which are easy to train on neural network based systems is usually done by a two step approach: segmentation followed by recognition. This approach has several shortcomings, which includes identification of text regions, layout diversity analysis present within pages and ground truth segmentation. These processes are prone to errors that create bottleneck in the recognition accuracies. Thus in this study, an end-to-end paragraph recognition system is presented with internal line segmentation and lexicon decoder as post processing step, which is free from those errors. We named our model as LexiconNet. In LexiconNet, given a paragraph image a combination of convolution and depth-wise separable convolutional modules generates the two dimension feature map of the image. The attention module is responsible for internal line segmentation that consequently processing a page in a line by line manner. At decoding step, we have added connectionist temporal classification based word beam search decoder as a post processing step. Our approach reports state-of-the-art results on standard datasets. The reported character error rate is 3.24% on IAM dataset with 27.19% improvement, 1.13% on RIMES with 40.83% improvement and 2.43% on READ-16 dataset with 32.31% improvement from existing literature and the word error rate is 8.29% on IAM dataset with 43.02% improvement, 2.94% on RIMES dataset with 56.25% improvement and 7.35% on READ-2016 dataset with 47.27% improvement from the existing results. The character error rate and word error rate reported in this work surpasses the results reported in literature.

55.MolMiner: You only look once for chemical structure recognition ⬇️

Molecular structures are always depicted as 2D printed form in scientific documents like journal papers and patents. However, these 2D depictions are not machine-readable. Due to a backlog of decades and an increasing amount of these printed literature, there is a high demand for the translation of printed depictions into machine-readable formats, which is known as Optical Chemical Structure Recognition (OCSR). Most OCSR systems developed over the last three decades follow a rule-based approach where the key step of vectorization of the depiction is based on the interpretation of vectors and nodes as bonds and atoms. Here, we present a practical software MolMiner, which is primarily built up using deep neural networks originally developed for semantic segmentation and object detection to recognize atom and bond elements from documents. These recognized elements can be easily connected as a molecular graph with distance-based construction algorithm. We carefully evaluate our software on four benchmark datasets with the state-of-the-art performance. Various real application scenarios are also tested, yielding satisfactory outcomes. The free download links of Mac and Windows versions are available: Mac: this https URL and Windows: this https URL

56.Boosting Multi-Label Image Classification with Complementary Parallel Self-Distillation ⬇️

Multi-Label Image Classification (MLIC) approaches usually exploit label correlations to achieve good performance. However, emphasizing correlation like co-occurrence may overlook discriminative features of the target itself and lead to model overfitting, thus undermining the performance. In this study, we propose a generic framework named Parallel Self-Distillation (PSD) for boosting MLIC models. PSD decomposes the original MLIC task into several simpler MLIC sub-tasks via two elaborated complementary task decomposition strategies named Co-occurrence Graph Partition (CGP) and Dis-occurrence Graph Partition (DGP). Then, the MLIC models of fewer categories are trained with these sub-tasks in parallel for respectively learning the joint patterns and the category-specific patterns of labels. Finally, knowledge distillation is leveraged to learn a compact global ensemble of full categories with these learned patterns for reconciling the label correlation exploitation and model overfitting. Extensive results on MS-COCO and NUS-WIDE datasets demonstrate that our framework can be easily plugged into many MLIC approaches and improve performances of recent state-of-the-art approaches. The explainable visual study also further validates that our method is able to learn both the category-specific and co-occurring features. The source code is released at this https URL.

57.Evaluating deep tracking models for player tracking in broadcast ice hockey video ⬇️

Tracking and identifying players is an important problem in computer vision based ice hockey analytics. Player tracking is a challenging problem since the motion of players in hockey is fast-paced and non-linear. There is also significant player-player and player-board occlusion, camera panning and zooming in hockey broadcast video. Prior published research perform player tracking with the help of handcrafted features for player detection and re-identification. Although commercial solutions for hockey player tracking exist, to the best of our knowledge, no network architectures used, training data or performance metrics are publicly reported. There is currently no published work for hockey player tracking making use of the recent advancements in deep learning while also reporting the current accuracy metrics used in literature. Therefore, in this paper, we compare and contrast several state-of-the-art tracking algorithms and analyze their performance and failure modes in ice hockey.

58.Visual Explanations from Deep Networks via Riemann-Stieltjes Integrated Gradient-based Localization ⬇️

Neural networks are becoming increasingly better at tasks that involve classifying and recognizing images. At the same time techniques intended to explain the network output have been proposed. One such technique is the Gradient-based Class Activation Map (Grad-CAM), which is able to locate features of an input image at various levels of a convolutional neural network (CNN), but is sensitive to the vanishing gradients problem. There are techniques such as Integrated Gradients (IG), that are not affected by that problem, but its use is limited to the input layer of a network. Here we introduce a new technique to produce visual explanations for the predictions of a CNN. Like Grad-CAM, our method can be applied to any layer of the network, and like Integrated Gradients it is not affected by the problem of vanishing gradients. For efficiency, gradient integration is performed numerically at the layer level using a Riemann-Stieltjes sum approximation. Compared to Grad-CAM, heatmaps produced by our algorithm are better focused in the areas of interest, and their numerical computation is more stable. Our code is available at this https URL

59.Geo-Localization via Ground-to-Satellite Cross-View Image Retrieval ⬇️

The large variation of viewpoint and irrelevant content around the target always hinder accurate image retrieval and its subsequent tasks. In this paper, we investigate an extremely challenging task: given a ground-view image of a landmark, we aim to achieve cross-view geo-localization by searching out its corresponding satellite-view images. Specifically, the challenge comes from the gap between ground-view and satellite-view, which includes not only large viewpoint changes (some parts of the landmark may be invisible from front view to top view) but also highly irrelevant background (the target landmark tend to be hidden in other surrounding buildings), making it difficult to learn a common representation or a suitable mapping.
To address this issue, we take advantage of drone-view information as a bridge between ground-view and satellite-view domains. We propose a Peer Learning and Cross Diffusion (PLCD) framework. PLCD consists of three parts: 1) a peer learning across ground-view and drone-view to find visible parts to benefit ground-drone cross-view representation learning; 2) a patch-based network for satellite-drone cross-view representation learning; 3) a cross diffusion between ground-drone space and satellite-drone space. Extensive experiments conducted on the University-Earth and University-Google datasets show that our method outperforms state-of-the-arts significantly.

60.Dynamic Query Selection for Fast Visual Perceiver ⬇️

Transformers have been matching deep convolutional networks for vision architectures in recent works. Most work is focused on getting the best results on large-scale benchmarks, and scaling laws seem to be the most successful strategy: bigger models, more data, and longer training result in higher performance. However, the reduction of network complexity and inference time remains under-explored. The Perceiver model offers a solution to this problem: by first performing a Cross-attention with a fixed number Q of latent query tokens, the complexity of the L-layers Transformer network that follows is bounded by O(LQ^2). In this work, we explore how to make Perceivers even more efficient, by reducing the number of queries Q during inference while limiting the accuracy drop.

61.Vision-based Anti-UAV Detection and Tracking ⬇️

Unmanned aerial vehicles (UAV) have been widely used in various fields, and their invasion of security and privacy has aroused social concern. Several detection and tracking systems for UAVs have been introduced in recent years, but most of them are based on radio frequency, radar, and other media. We assume that the field of computer vision is mature enough to detect and track invading UAVs. Thus we propose a visible light mode dataset called Dalian University of Technology Anti-UAV dataset, DUT Anti-UAV for short. It contains a detection dataset with a total of 10,000 images and a tracking dataset with 20 videos that include short-term and long-term sequences. All frames and images are manually annotated precisely. We use this dataset to train several existing detection algorithms and evaluate the algorithms' performance. Several tracking methods are also tested on our tracking dataset. Furthermore, we propose a clear and simple tracking algorithm combined with detection that inherits the detector's high precision. Extensive experiments show that the tracking performance is improved considerably after fusing detection, thus providing a new attempt at UAV tracking using our dataset.The datasets and results are publicly available at: this https URL

62.Self-supervised U-net for few-shot learning of object segmentation in microscopy images ⬇️

State-of-the-art segmentation performances are achieved by deep neural networks. Training these networks from only a few training examples is challenging while producing annotated images that provide supervision is tedious. Recently, self-supervision, i.e. designing a neural pipeline providing synthetic or indirect supervision, has proved to significantly increase generalization performances of models trained on few shots. This paper introduces one such neural pipeline in the context of microscopic image segmentation. By leveraging the rather simple content of these images a trainee network can be mentored by a referee network which has been previously trained on synthetically generated pairs of corrupted/correct region masks.

63.Deep Learning for Visual Speech Analysis: A Survey ⬇️

Visual speech, referring to the visual domain of speech, has attracted increasing attention due to its wide applications, such as public security, medical treatment, military defense, and film entertainment. As a powerful AI strategy, deep learning techniques have extensively promoted the development of visual speech learning. Over the past five years, numerous deep learning based methods have been proposed to address various problems in this area, especially automatic visual speech recognition and generation. To push forward future research on visual speech, this paper aims to present a comprehensive review of recent progress in deep learning methods on visual speech analysis. We cover different aspects of visual speech, including fundamental problems, challenges, benchmark datasets, a taxonomy of existing methods, and state-of-the-art performance. Besides, we also identify gaps in current research and discuss inspiring future research directions.

64.Grad-CAM++ is Equivalent to Grad-CAM With Positive Gradients ⬇️

The Grad-CAM algorithm provides a way to identify what parts of an image contribute most to the output of a classifier deep network. The algorithm is simple and widely used for localization of objects in an image, although some researchers have point out its limitations, and proposed various alternatives. One of them is Grad-CAM++, that according to its authors can provide better visual explanations for network predictions, and does a better job at locating objects even for occurrences of multiple object instances in a single image. Here we show that Grad-CAM++ is practically equivalent to a very simple variation of Grad-CAM in which gradients are replaced with positive gradients.

65.ReLU Fields: The Little Non-linearity That Could ⬇️

In many recent works, multi-layer perceptions (MLPs) have been shown to be suitable for modeling complex spatially-varying functions including images and 3D scenes. Although the MLPs are able to represent complex scenes with unprecedented quality and memory footprint, this expressive power of the MLPs, however, comes at the cost of long training and inference times. On the other hand, bilinear/trilinear interpolation on regular grid based representations can give fast training and inference times, but cannot match the quality of MLPs without requiring significant additional memory. Hence, in this work, we investigate what is the smallest change to grid-based representations that allows for retaining the high fidelity result of MLPs while enabling fast reconstruction and rendering times. We introduce a surprisingly simple change that achieves this task -- simply allowing a fixed non-linearity (ReLU) on interpolated grid values. When combined with coarse to-fine optimization, we show that such an approach becomes competitive with the state-of-the-art. We report results on radiance fields, and occupancy fields, and compare against multiple existing alternatives. Code and data for the paper are available at this https URL.

66.Knowledge Distillation via the Target-aware Transformer ⬇️

Knowledge distillation becomes a de facto standard to improve the performance of small neural networks. Most of the previous works propose to regress the representational features from the teacher to the student in a one-to-one spatial matching fashion. However, people tend to overlook the fact that, due to the architecture differences, the semantic information on the same spatial location usually vary. This greatly undermines the underlying assumption of the one-to-one distillation approach. To this end, we propose a novel one-to-all spatial matching knowledge distillation approach. Specifically, we allow each pixel of the teacher feature to be distilled to all spatial locations of the student features given its similarity, which is generated from a target-aware transformer. Our approach surpasses the state-of-the-art methods by a significant margin on various computer vision benchmarks, such as ImageNet, Pascal VOC and COCOStuff10k. Code will be released soon.

67.Learning Muti-expert Distribution Calibration for Long-tailed Video Classification ⬇️

Most existing state-of-the-art video classification methods assume the training data obey a uniform distribution. However, video data in the real world typically exhibit long-tail class distribution and imbalance, which extensively results in a model bias towards head class and leads to relatively low performance on tail class. While the current long-tail classification methods usually focus on image classification, adapting it to video data is not a trivial extension. We propose an end-to-end multi-experts distribution calibration method based on two-level distribution information to address these challenges. The method jointly considers the distribution of samples in each class (intra-class distribution) and the diverse distributions of overall data (inter-class distribution) to solve the problem of imbalanced data under long-tailed distribution. By modeling this two-level distribution information, the model can consider the head classes and the tail classes and significantly transfer the knowledge from the head classes to improve the performance of the tail classes. Extensive experiments verify that our method achieves state-of-the-art performance on the long-tailed video classification task.

68.Human Instance Matting via Mutual Guidance and Multi-Instance Refinement ⬇️

This paper introduces a new matting task called human instance matting (HIM), which requires the pertinent model to automatically predict a precise alpha matte for each human instance. Straightforward combination of closely related techniques, namely, instance segmentation, soft segmentation and human/conventional matting, will easily fail in complex cases requiring disentangling mingled colors belonging to multiple instances along hairy and thin boundary structures. To tackle these technical challenges, we propose a human instance matting framework, called InstMatt, where a novel mutual guidance strategy working in tandem with a multi-instance refinement module is used, for delineating multi-instance relationship among humans with complex and overlapping boundaries if present. A new instance matting metric called instance matting quality (IMQ) is proposed, which addresses the absence of a unified and fair means of evaluation emphasizing both instance recognition and matting quality. Finally, we construct a HIM benchmark for evaluation, which comprises of both synthetic and natural benchmark images. In addition to thorough experimental results on complex cases with multiple and overlapping human instances each has intricate boundaries, preliminary results are presented on general instance matting. Code and benchmark are available in this https URL.

69.Recent Advances in Embedding Methods for Multi-Object Tracking: A Survey ⬇️

Multi-object tracking (MOT) aims to associate target objects across video frames in order to obtain entire moving trajectories. With the advancement of deep neural networks and the increasing demand for intelligent video analysis, MOT has gained significantly increased interest in the computer vision community. Embedding methods play an essential role in object location estimation and temporal identity association in MOT. Unlike other computer vision tasks, such as image classification, object detection, re-identification, and segmentation, embedding methods in MOT have large variations, and they have never been systematically analyzed and summarized. In this survey, we first conduct a comprehensive overview with in-depth analysis for embedding methods in MOT from seven different perspectives, including patch-level embedding, single-frame embedding, cross-frame joint embedding, correlation embedding, sequential embedding, tracklet embedding, and cross-track relational embedding. We further summarize the existing widely used MOT datasets and analyze the advantages of existing state-of-the-art methods according to their embedding strategies. Finally, some critical yet under-investigated areas and future research directions are discussed.

70.Evidence for Hypodescent in Visual Semantic AI ⬇️

We examine the state-of-the-art multimodal "visual semantic" model CLIP ("Contrastive Language Image Pretraining") for the rule of hypodescent, or one-drop rule, whereby multiracial people are more likely to be assigned a racial or ethnic label corresponding to a minority or disadvantaged racial or ethnic group than to the equivalent majority or advantaged group. A face morphing experiment grounded in psychological research demonstrating hypodescent indicates that, at the midway point of 1,000 series of morphed images, CLIP associates 69.7% of Black-White female images with a Black text label over a White text label, and similarly prefers Latina (75.8%) and Asian (89.1%) text labels at the midway point for Latina-White female and Asian-White female morphs, reflecting hypodescent. Additionally, assessment of the underlying cosine similarities in the model reveals that association with White is correlated with association with "person," with Pearson's rho as high as 0.82 over a 21,000-image morph series, indicating that a White person corresponds to the default representation of a person in CLIP. Finally, we show that the stereotype-congruent pleasantness association of an image correlates with association with the Black text label in CLIP, with Pearson's rho = 0.48 for 21,000 Black-White multiracial male images, and rho = 0.41 for Black-White multiracial female images. CLIP is trained on English-language text gathered using data collected from an American website (Wikipedia), and our findings demonstrate that CLIP embeds the values of American racial hierarchy, reflecting the implicit and explicit beliefs that are present in human minds. We contextualize these findings within the history and psychology of hypodescent. Overall, the data suggests that AI supervised using natural language will, unless checked, learn biases that reflect racial hierarchies.

71.CNNs are Myopic ⬇️

We claim that Convolutional Neural Networks (CNNs) learn to classify images using only small seemingly unrecognizable tiles. We show experimentally that CNNs trained only using such tiles can match or even surpass the performance of CNNs trained on full images. Conversely, CNNs trained on full images show similar predictions on small tiles. We also propose the first a priori theoretical model for convolutional data sets that seems to explain this behavior. This gives additional support to the long standing suspicion that CNNs do not need to understand the global structure of images to achieve state-of-the-art accuracies. Surprisingly it also suggests that over-fitting is not needed either.

72.Real Time Detection Free Tracking of Multiple Objects Via Equilibrium Optimizer ⬇️

Multiple objects tracking (MOT) is a difficult task, as it usually requires special hardware and higher computation complexity. In this work, we present a new framework of MOT by using of equilibrium optimizer (EO) algorithm and reducing the resolution of the bounding boxes of the objects to solve such problems in the detection free framework. First, in the first frame the target objects are initialized and its size is computed, then its resolution is reduced if it is higher than a threshold, and then modeled by their kernel color histogram to establish a feature model. The Bhattacharya distances between the histogram of object models and other candidates are used as the fitness function to be optimized. Multiple agents are generated by EO, according to the number of the target objects to be tracked. EO algorithm is used because of its efficiency and lower computation cost compared to other algorithms in global optimization. Experimental results confirm that EO multi-object tracker achieves satisfying tracking results then other trackers.

73.Language Models with Image Descriptors are Strong Few-Shot Video-Language Learners ⬇️

The goal of this work is to build flexible video-language models that can generalize to various video-to-text tasks from few examples, such as domain-specific captioning, question answering, and future event prediction. Existing few-shot video-language learners focus exclusively on the encoder, resulting in the absence of a video-to-text decoder to handle generative tasks. Video captioners have been pretrained on large-scale video-language datasets, but they rely heavily on finetuning and lack the ability to generate text for unseen tasks in a few-shot setting. We propose VidIL, a few-shot Video-language Learner via Image and Language models, which demonstrates strong performance on few-shot video-to-text tasks without the necessity of pretraining or finetuning on any video datasets. We use the image-language models to translate the video content into frame captions, object, attribute, and event phrases, and compose them into a temporal structure template. We then instruct a language model, with a prompt containing a few in-context examples, to generate a target output from the composed content. The flexibility of prompting allows the model to capture any form of text input, such as automatic speech recognition (ASR) transcripts. Our experiments demonstrate the power of language models in understanding videos on a wide variety of video-language tasks, including video captioning, video question answering, video caption retrieval, and video future event prediction. Especially, on video future event prediction, our few-shot model significantly outperforms state-of-the-art supervised models trained on large-scale video datasets. Code and resources are publicly available for research purposes at this https URL .

74.Classification of Quasars, Galaxies, and Stars in the Mapping of the Universe Multi-modal Deep Learning ⬇️

In this paper, the fourth version the Sloan Digital Sky Survey (SDSS-4), Data Release 16 dataset was used to classify the SDSS dataset into galaxies, stars, and quasars using machine learning and deep learning architectures. We efficiently utilize both image and metadata in tabular format to build a novel multi-modal architecture and achieve state-of-the-art results. In addition, our experiments on transfer learning using Imagenet weights on five different architectures (Resnet-50, DenseNet-121 VGG-16, Xception, and EfficientNet) reveal that freezing all layers and adding a final trainable layer may not be an optimal solution for transfer learning. It is hypothesized that higher the number of trainable layers, higher will be the training time and accuracy of predictions. It is also hypothesized that any subsequent increase in the number of training layers towards the base layers will not increase in accuracy as the pre trained lower layers only help in low level feature extraction which would be quite similar in all the datasets. Hence the ideal level of trainable layers needs to be identified for each model in respect to the number of parameters. For the tabular data, we compared classical machine learning algorithms (Logistic Regression, Random Forest, Decision Trees, Adaboost, LightGBM etc.,) with artificial neural networks. Our works shed new light on transfer learning and multi-modal deep learning architectures. The multi-modal architecture not only resulted in higher metrics (accuracy, precision, recall, F1 score) than models using only image data or tabular data. Furthermore, multi-modal architecture achieved the best metrics in lesser training epochs and improved the metrics on all classes.

75.OTAdapt: Optimal Transport-based Approach For Unsupervised Domain Adaptation ⬇️

Unsupervised domain adaptation is one of the challenging problems in computer vision. This paper presents a novel approach to unsupervised domain adaptations based on the optimal transport-based distance. Our approach allows aligning target and source domains without the requirement of meaningful metrics across domains. In addition, the proposal can associate the correct mapping between source and target domains and guarantee a constraint of topology between source and target domains. The proposed method is evaluated on different datasets in various problems, i.e. (i) digit recognition on MNIST, MNIST-M, USPS datasets, (ii) Object recognition on Amazon, Webcam, DSLR, and VisDA datasets, (iii) Insect Recognition on the IP102 dataset. The experimental results show that our proposed method consistently improves performance accuracy. Also, our framework could be incorporated with any other CNN frameworks within an end-to-end deep network design for recognition problems to improve their performance.

76.Learnable Visual Words for Interpretable Image Recognition ⬇️

To interpret deep models' predictions, attention-based visual cues are widely used in addressing \textit{why} deep models make such predictions. Beyond that, the current research community becomes more interested in reasoning \textit{how} deep models make predictions, where some prototype-based methods employ interpretable representations with their corresponding visual cues to reveal the black-box mechanism of deep model behaviors. However, these pioneering attempts only either learn the category-specific prototypes and deteriorate their generalizing capacities, or demonstrate several illustrative examples without a quantitative evaluation of visual-based interpretability with further limitations on their practical usages. In this paper, we revisit the concept of visual words and propose the Learnable Visual Words (LVW) to interpret the model prediction behaviors with two novel modules: semantic visual words learning and dual fidelity preservation. The semantic visual words learning relaxes the category-specific constraint, enabling the general visual words shared across different categories. Beyond employing the visual words for prediction to align visual words with the base model, our dual fidelity preservation also includes the attention guided semantic alignment that encourages the learned visual words to focus on the same conceptual regions for prediction. Experiments on six visual benchmarks demonstrate the superior effectiveness of our proposed LVW in both accuracy and model interpretation over the state-of-the-art methods. Moreover, we elaborate on various in-depth analyses to further explore the learned visual words and the generalizability of our method for unseen categories.

77.Housekeep: Tidying Virtual Households using Commonsense Reasoning ⬇️

We introduce Housekeep, a benchmark to evaluate commonsense reasoning in the home for embodied AI. In Housekeep, an embodied agent must tidy a house by rearranging misplaced objects without explicit instructions specifying which objects need to be rearranged. Instead, the agent must learn from and is evaluated against human preferences of which objects belong where in a tidy house. Specifically, we collect a dataset of where humans typically place objects in tidy and untidy houses constituting 1799 objects, 268 object categories, 585 placements, and 105 rooms. Next, we propose a modular baseline approach for Housekeep that integrates planning, exploration, and navigation. It leverages a fine-tuned large language model (LLM) trained on an internet text corpus for effective planning. We show that our baseline agent generalizes to rearranging unseen objects in unknown environments. See our webpage for more details: this https URL

78.GL-RG: Global-Local Representation Granularity for Video Captioning ⬇️

Video captioning is a challenging task as it needs to accurately transform visual understanding into natural language description. To date, state-of-the-art methods inadequately model global-local representation across video frames for caption generation, leaving plenty of room for improvement. In this work, we approach the video captioning task from a new perspective and propose a GL-RG framework for video captioning, namely a \textbf{G}lobal-\textbf{L}ocal \textbf{R}epresentation \textbf{G}ranularity. Our GL-RG demonstrates three advantages over the prior efforts: 1) we explicitly exploit extensive visual representations from different video ranges to improve linguistic expression; 2) we devise a novel global-local encoder to produce rich semantic vocabulary to obtain a descriptive granularity of video contents across frames; 3) we develop an incremental training strategy which organizes model learning in an incremental fashion to incur an optimal captioning behavior. Experimental results on the challenging MSR-VTT and MSVD datasets show that our DL-RG outperforms recent state-of-the-art methods by a significant margin. Code is available at \url{this https URL}.

79.Individual Topology Structure of Eye Movement Trajectories ⬇️

Traditionally, extracting patterns from eye movement data relies on statistics of different macro-events such as fixations and saccades. This requires an additional preprocessing step to separate the eye movement subtypes, often with a number of parameters on which the classification results depend. Besides that, definitions of such macro events are formulated in different ways by different researchers.
We propose an application of a new class of features to the quantitative analysis of personal eye movement trajectories structure. This new class of features based on algebraic topology allows extracting patterns from different modalities of gaze such as time series of coordinates and amplitudes, heatmaps, and point clouds in a unified way at all scales from micro to macro. We experimentally demonstrate the competitiveness of the new class of features with the traditional ones and their significant synergy while being used together for the person authentication task on the recently published eye movement trajectories dataset.

80.Vision Transformers in 2022: An Update on Tiny ImageNet ⬇️

The recent advances in image transformers have shown impressive results and have largely closed the gap between traditional CNN architectures. The standard procedure is to train on large datasets like ImageNet-21k and then finetune on ImageNet-1k. After finetuning, researches will often consider the transfer learning performance on smaller datasets such as CIFAR-10/100 but have left out Tiny ImageNet. This paper offers an update on vision transformers' performance on Tiny ImageNet. I include Vision Transformer (ViT) , Data Efficient Image Transformer (DeiT), Class Attention in Image Transformer (CaiT), and Swin Transformers. In addition, Swin Transformers beats the current state-of-the-art result with a validation accuracy of 91.35%. Code is available here: this https URL

81.Swept-Angle Synthetic Wavelength Interferometry ⬇️

We present a new imaging technique, swept-angle synthetic wavelength interferometry, for full-field micron-scale 3D sensing. As in conventional synthetic wavelength interferometry, our technique uses light consisting of two optical wavelengths, resulting in per-pixel interferometric measurements whose phase encodes scene depth. Our technique additionally uses a new type of light source that, by emulating spatially-incoherent illumination, makes interferometric measurements insensitive to global illumination effects that confound depth information. The resulting technique combines the speed of full-field interferometric setups with the robustness to global illumination of scanning interferometric setups. Overall, our technique can recover full-frame depth at a spatial and axial resolution of a few micrometers using as few as 16 measurements, resulting in fast acquisition at frame rates of 10 Hz. We build an experimental prototype and use it to demonstrate these capabilities, by scanning a variety of scenes that contain challenging light transport effects such as interreflections, subsurface scattering, and specularities. We validate the accuracy of our measurements by showing that they closely match reference measurements from a full-field optical coherence tomography system, despite being captured at orders of magnitude faster acquisition times and while operating under strong ambient light.

82.Towards real-time and energy efficient Siamese tracking -- a hardware-software approach ⬇️

Siamese trackers have been among the state-of-the-art solutions in each Visual Object Tracking (VOT) challenge over the past few years. However, with great accuracy comes great computational complexity: to achieve real-time processing, these trackers have to be massively parallelised and are usually run on high-end GPUs. Easy to implement, this approach is energy consuming, and thus cannot be used in many low-power applications. To overcome this, one can use energy-efficient embedded devices, such as heterogeneous platforms joining the ARM processor system with programmable logic (FPGA). In this work, we propose a hardware-software implementation of the well-known fully connected Siamese tracker (SiamFC). We have developed a quantised Siamese network for the FINN accelerator, using algorithm-accelerator co-design, and performed design space exploration to achieve the best efficiency-to-energy ratio (determined by FPS and used resources). For our network, running in the programmable logic part of the Zynq UltraScale+ MPSoC ZCU104, we achieved the processing of almost 50 frames-per-second with tracker accuracy on par with its floating point counterpart, as well as the original SiamFC network. The complete tracking system, implemented in ARM with the network accelerated on FPGA, achieves up to 17 fps. These results bring us towards bridging the gap between the highly accurate but energy-demanding algorithms and energy-efficient solutions ready to be used in low-power, edge systems.

83.Transformer-based out-of-distribution detection for clinically safe segmentation ⬇️

In a clinical setting it is essential that deployed image processing systems are robust to the full range of inputs they might encounter and, in particular, do not make confidently wrong predictions. The most popular approach to safe processing is to train networks that can provide a measure of their uncertainty, but these tend to fail for inputs that are far outside the training data distribution. Recently, generative modelling approaches have been proposed as an alternative; these can quantify the likelihood of a data sample explicitly, filtering out any out-of-distribution (OOD) samples before further processing is performed. In this work, we focus on image segmentation and evaluate several approaches to network uncertainty in the far-OOD and near-OOD cases for the task of segmenting haemorrhages in head CTs. We find all of these approaches are unsuitable for safe segmentation as they provide confidently wrong predictions when operating OOD. We propose performing full 3D OOD detection using a VQ-GAN to provide a compressed latent representation of the image and a transformer to estimate the data likelihood. Our approach successfully identifies images in both the far- and near-OOD cases. We find a strong relationship between image likelihood and the quality of a model's segmentation, making this approach viable for filtering images unsuitable for segmentation. To our knowledge, this is the first time transformers have been applied to perform OOD detection on 3D image data.

84.AutoLink: Self-supervised Learning of Human Skeletons and Object Outlines by Linking Keypoints ⬇️

Structured representations such as keypoints are widely used in pose transfer, conditional image generation, animation, and 3D reconstruction. However, their supervised learning requires expensive annotation for each target domain. We propose a self-supervised method that learns to disentangle object structure from the appearance with a graph of 2D keypoints linked by straight edges. Both the keypoint location and their pairwise edge weights are learned, given only a collection of images depicting the same object class. The graph is interpretable, for example, AutoLink recovers the human skeleton topology when applied to images showing people. Our key ingredients are i) an encoder that predicts keypoint locations in an input image, ii) a shared graph as a latent variable that links the same pairs of keypoints in every image, iii) an intermediate edge map that combines the latent graph edge weights and keypoint locations in a soft, differentiable manner, and iv) an inpainting objective on randomly masked images. Although simpler, AutoLink outperforms existing self-supervised methods on the established keypoint and pose estimation benchmarks and paves the way for structure-conditioned generative models on more diverse datasets.

85.Gradient Concealment: Free Lunch for Defending Adversarial Attacks ⬇️

Recent studies show that the deep neural networks (DNNs) have achieved great success in various tasks. However, even the \emph{state-of-the-art} deep learning based classifiers are extremely vulnerable to adversarial examples, resulting in sharp decay of discrimination accuracy in the presence of enormous unknown attacks. Given the fact that neural networks are widely used in the open world scenario which can be safety-critical situations, mitigating the adversarial effects of deep learning methods has become an urgent need. Generally, conventional DNNs can be attacked with a dramatically high success rate since their gradient is exposed thoroughly in the white-box scenario, making it effortless to ruin a well trained classifier with only imperceptible perturbations in the raw data space. For tackling this problem, we propose a plug-and-play layer that is training-free, termed as \textbf{G}radient \textbf{C}oncealment \textbf{M}odule (GCM), concealing the vulnerable direction of gradient while guaranteeing the classification accuracy during the inference time. GCM reports superior defense results on the ImageNet classification benchmark, improving up to 63.41% top-1 attack robustness (AR) when faced with adversarial inputs compared to the vanilla DNNs. Moreover, we use GCM in the CVPR 2022 Robust Classification Challenge, currently achieving \textbf{2nd} place in Phase II with only a tiny version of ConvNext. The code will be made available.

86.Lightweight Human Pose Estimation Using Heatmap-Weighting Loss ⬇️

Recent research on human pose estimation exploits complex structures to improve performance on benchmark datasets, ignoring the resource overhead and inference speed when the model is actually deployed. In this paper, we lighten the computation cost and parameters of the deconvolution head network in SimpleBaseline and introduce an attention mechanism that utilizes original, inter-level, and intra-level information to intensify the accuracy. Additionally, we propose a novel loss function called heatmap weighting loss, which generates weights for each pixel on the heatmap that makes the model more focused on keypoints. Experiments demonstrate our method achieves a balance between performance, resource volume, and inference speed. Specifically, our method can achieve 65.3 AP score on COCO test-dev, while the inference speed is 55 FPS and 18 FPS on the mobile GPU and CPU, respectively.

87.Facing the Void: Overcoming Missing Data in Multi-View Imagery ⬇️

In some scenarios, a single input image may not be enough to allow the object classification. In those cases, it is crucial to explore the complementary information extracted from images presenting the same object from multiple perspectives (or views) in order to enhance the general scene understanding and, consequently, increase the performance. However, this task, commonly called multi-view image classification, has a major challenge: missing data. In this paper, we propose a novel technique for multi-view image classification robust to this problem. The proposed method, based on state-of-the-art deep learning-based approaches and metric learning, can be easily adapted and exploited in other applications and domains. A systematic evaluation of the proposed algorithm was conducted using two multi-view aerial-ground datasets with very distinct properties. Results show that the proposed algorithm provides improvements in multi-view image classification accuracy when compared to state-of-the-art methods. Code available at \url{this https URL}.

88.A comprehensive survey on semantic facial attribute editing using generative adversarial networks ⬇️

Generating random photo-realistic images has experienced tremendous growth during the past few years due to the advances of the deep convolutional neural networks and generative models. Among different domains, face photos have received a great deal of attention and a large number of face generation and manipulation models have been proposed. Semantic facial attribute editing is the process of varying the values of one or more attributes of a face image while the other attributes of the image are not affected. The requested modifications are provided as an attribute vector or in the form of driving face image and the whole process is performed by the corresponding models. In this paper, we survey the recent works and advances in semantic facial attribute editing. We cover all related aspects of these models including the related definitions and concepts, architectures, loss functions, datasets, evaluation metrics, and applications. Based on their architectures, the state-of-the-art models are categorized and studied as encoder-decoder, image-to-image, and photo-guided models. The challenges and restrictions of the current state-of-the-art methods are discussed as well.

89.Boosting Camouflaged Object Detection with Dual-Task Interactive Transformer ⬇️

Camouflaged object detection intends to discover the concealed objects hidden in the surroundings. Existing methods follow the bio-inspired framework, which first locates the object and second refines the boundary. We argue that the discovery of camouflaged objects depends on the recurrent search for the object and the boundary. The recurrent processing makes the human tired and helpless, but it is just the advantage of the transformer with global search ability. Therefore, a dual-task interactive transformer is proposed to detect both accurate position of the camouflaged object and its detailed boundary. The boundary feature is considered as Query to improve the camouflaged object detection, and meanwhile the object feature is considered as Query to improve the boundary detection. The camouflaged object detection and the boundary detection are fully interacted by multi-head self-attention. Besides, to obtain the initial object feature and boundary feature, transformer-based backbones are adopted to extract the foreground and background. The foreground is just object, while foreground minus background is considered as boundary. Here, the boundary feature can be obtained from blurry boundary region of the foreground and background. Supervised by the object, the background and the boundary ground truth, the proposed model achieves state-of-the-art performance in public datasets. this https URL

90.Multi-feature Co-learning for Image Inpainting ⬇️

Image inpainting has achieved great advances by simultaneously leveraging image structure and texture features. However, due to lack of effective multi-feature fusion techniques, existing image inpainting methods still show limited improvement. In this paper, we design a deep multi-feature co-learning network for image inpainting, which includes Soft-gating Dual Feature Fusion (SDFF) and Bilateral Propagation Feature Aggregation (BPFA) modules. To be specific, we first use two branches to learn structure features and texture features separately. Then the proposed SDFF module integrates structure features into texture features, and meanwhile uses texture features as an auxiliary in generating structure features. Such a co-learning strategy makes the structure and texture features more consistent. Next, the proposed BPFA module enhances the connection from local feature to overall consistency by co-learning contextual attention, channel-wise information and feature space, which can further refine the generated structures and textures. Finally, extensive experiments are performed on benchmark datasets, including CelebA, Places2, and Paris StreetView. Experimental results demonstrate the superiority of the proposed method over the state-of-the-art. The source codes are available at this https URL.

91.ADT-SSL: Adaptive Dual-Threshold for Semi-Supervised Learning ⬇️

Semi-Supervised Learning (SSL) has advanced classification tasks by inputting both labeled and unlabeled data to train a model jointly. However, existing SSL methods only consider the unlabeled data whose predictions are beyond a fixed threshold (e.g., 0.95), ignoring the valuable information from those less than 0.95. We argue that these discarded data have a large proportion and are usually of hard samples, thereby benefiting the model training. This paper proposes an Adaptive Dual-Threshold method for Semi-Supervised Learning (ADT-SSL). Except for the fixed threshold, ADT extracts another class-adaptive threshold from the labeled data to take full advantage of the unlabeled data whose predictions are less than 0.95 but more than the extracted one. Accordingly, we engage CE and $L_2$ loss functions to learn from these two types of unlabeled data, respectively. For highly similar unlabeled data, we further design a novel similar loss to make the prediction of the model consistency. Extensive experiments are conducted on benchmark datasets, including CIFAR-10, CIFAR-100, and SVHN. Experimental results show that the proposed ADT-SSL achieves state-of-the-art classification accuracy.

92.Cycle-GAN for eye-tracking ⬇️

This manuscript presents a not typical implementation of the cycle generative adversarial networks (Cycle-GAN) method for eye-tracking tasks.

93.Robot Person Following in Uniform Crowd Environment ⬇️

Person-tracking robots have many applications, such as in security, elderly care, and socializing robots. Such a task is particularly challenging when the person is moving in a Uniform crowd. Also, despite significant progress of trackers reported in the literature, state-of-the-art trackers have hardly addressed person following in such scenarios. In this work, we focus on improving the perceptivity of a robot for a person following task by developing a robust and real-time applicable object tracker. We present a new robot person tracking system with a new RGB-D tracker, Deep Tracking with RGB-D (DTRD) that is resilient to tricky challenges introduced by the uniform crowd environment. Our tracker utilizes transformer encoder-decoder architecture with RGB and depth information to discriminate the target person from similar distractors. A substantial amount of comprehensive experiments and results demonstrate that our tracker has higher performance in two quantitative evaluation metrics and confirms its superiority over other SOTA trackers.

94.Improvements to Self-Supervised Representation Learning for Masked Image Modeling ⬇️

This paper explores improvements to the masked image modeling (MIM) paradigm. The MIM paradigm enables the model to learn the main object features of the image by masking the input image and predicting the masked part by the unmasked part. We found the following three main directions for MIM to be improved. First, since both encoders and decoders contribute to representation learning, MIM uses only encoders for downstream tasks, which ignores the impact of decoders on representation learning. Although the MIM paradigm already employs small decoders with asymmetric structures, we believe that continued reduction of decoder parameters is beneficial to improve the representational learning capability of the encoder . Second, MIM solves the image prediction task by training the encoder and decoder together , and does not design a separate task for the encoder . To further enhance the performance of the encoder when performing downstream tasks, we designed the encoder for the tasks of comparative learning and token position prediction. Third, since the input image may contain background and other objects, and the proportion of each object in the image varies, reconstructing the tokens related to the background or to other objects is not meaningful for MIM to understand the main object representations. Therefore we use ContrastiveCrop to crop the input image so that the input image contains as much as possible only the main objects. Based on the above three improvements to MIM, we propose a new model, Contrastive Masked AutoEncoders (CMAE). We achieved a Top-1 accuracy of 65.84% on tinyimagenet using the ViT-B backbone, which is +2.89 outperforming the MAE of competing methods when all conditions are equal. Code will be made available.

95.On the Feasibility and Generality of Patch-based Adversarial Attacks on Semantic Segmentation Problems ⬇️

Deep neural networks were applied with success in a myriad of applications, but in safety critical use cases adversarial attacks still pose a significant threat. These attacks were demonstrated on various classification and detection tasks and are usually considered general in a sense that arbitrary network outputs can be generated by them.
In this paper we will demonstrate through simple case studies both in simulation and in real-life, that patch based attacks can be utilised to alter the output of segmentation networks. Through a few examples and the investigation of network complexity, we will also demonstrate that the number of possible output maps which can be generated via patch-based attacks of a given size is typically smaller than the area they effect or areas which should be attacked in case of practical applications.
We will prove that based on these results most patch-based attacks cannot be general in practice, namely they can not generate arbitrary output maps or if they could, they are spatially limited and this limit is significantly smaller than the receptive field of the patches.

96.Knowledge Distillation from A Stronger Teacher ⬇️

Unlike existing knowledge distillation methods focus on the baseline settings, where the teacher models and training strategies are not that strong and competing as state-of-the-art approaches, this paper presents a method dubbed DIST to distill better from a stronger teacher. We empirically find that the discrepancy of predictions between the student and a stronger teacher may tend to be fairly severer. As a result, the exact match of predictions in KL divergence would disturb the training and make existing methods perform poorly. In this paper, we show that simply preserving the relations between the predictions of teacher and student would suffice, and propose a correlation-based loss to capture the intrinsic inter-class relations from the teacher explicitly. Besides, considering that different instances have different semantic similarities to each class, we also extend this relational match to the intra-class level. Our method is simple yet practical, and extensive experiments demonstrate that it adapts well to various architectures, model sizes and training strategies, and can achieve state-of-the-art performance consistently on image classification, object detection, and semantic segmentation tasks. Code is available at: this https URL .

97.Fine-Grained Visual Classification using Self Assessment Classifier ⬇️

Extracting discriminative features plays a crucial role in the fine-grained visual classification task. Most of the existing methods focus on developing attention or augmentation mechanisms to achieve this goal. However, addressing the ambiguity in the top-k prediction classes is not fully investigated. In this paper, we introduce a Self Assessment Classifier, which simultaneously leverages the representation of the image and top-k prediction classes to reassess the classification results. Our method is inspired by continual learning with coarse-grained and fine-grained classifiers to increase the discrimination of features in the backbone and produce attention maps of informative areas on the image. In practice, our method works as an auxiliary branch and can be easily integrated into different architectures. We show that by effectively addressing the ambiguity in the top-k prediction classes, our method achieves new state-of-the-art results on CUB200-2011, Stanford Dog, and FGVC Aircraft datasets. Furthermore, our method also consistently improves the accuracy of different existing fine-grained classifiers with a unified setup.

98.Point is a Vector: A Feature Representation in Point Analysis ⬇️

The irregularity and disorder of point clouds bring many challenges to point cloud analysis. PointMLP suggests that geometric information is not the only critical point in point cloud analysis. It achieves promising result based on a simple multi-layer perception (MLP) structure with geometric affine module. However, these MLP-like structures aggregate features only with fixed weights, while differences in the semantic information of different point features are ignored. So we propose a novel Point-Vector Representation of the point feature to improve feature aggregation by using inductive bias. The direction of the introduced vector representation can dynamically modulate the aggregation of two point features according to the semantic relationship. Based on it, we design a novel Point2Vector MLP architecture. Experiments show that it achieves state-of-the-art performance on the classification task of ScanObjectNN dataset, with 1% increase, compared with the previous best method. We hope our method can help people better understand the role of semantic information in point cloud analysis and lead to explore more and better feature representations or other ways.

99.Semi-Supervised Subspace Clustering via Tensor Low-Rank Representation ⬇️

In this letter, we propose a novel semi-supervised subspace clustering method, which is able to simultaneously augment the initial supervisory information and construct a discriminative affinity matrix. By representing the limited amount of supervisory information as a pairwise constraint matrix, we observe that the ideal affinity matrix for clustering shares the same low-rank structure as the ideal pairwise constraint matrix. Thus, we stack the two matrices into a 3-D tensor, where a global low-rank constraint is imposed to promote the affinity matrix construction and augment the initial pairwise constraints synchronously. Besides, we use the local geometry structure of input samples to complement the global low-rank prior to achieve better affinity matrix learning. The proposed model is formulated as a Laplacian graph regularized convex low-rank tensor representation problem, which is further solved with an alternative iterative algorithm. In addition, we propose to refine the affinity matrix with the augmented pairwise constraints. Comprehensive experimental results on six commonly-used benchmark datasets demonstrate the superiority of our method over state-of-the-art methods. The code is publicly available at this https URL.

100.Deep Learning for Omnidirectional Vision: A Survey and New Perspectives ⬇️

Omnidirectional image (ODI) data is captured with a 360x180 field-of-view, which is much wider than the pinhole cameras and contains richer spatial information than the conventional planar images. Accordingly, omnidirectional vision has attracted booming attention due to its more advantageous performance in numerous applications, such as autonomous driving and virtual reality. In recent years, the availability of customer-level 360 cameras has made omnidirectional vision more popular, and the advance of deep learning (DL) has significantly sparked its research and applications. This paper presents a systematic and comprehensive review and analysis of the recent progress in DL methods for omnidirectional vision. Our work covers four main contents: (i) An introduction to the principle of omnidirectional imaging, the convolution methods on the ODI, and datasets to highlight the differences and difficulties compared with the 2D planar image data; (ii) A structural and hierarchical taxonomy of the DL methods for omnidirectional vision; (iii) A summarization of the latest novel learning strategies and applications; (iv) An insightful discussion of the challenges and open problems by highlighting the potential research directions to trigger more research in the community.

101.Robust Sensible Adversarial Learning of Deep Neural Networks for Image Classification ⬇️

The idea of robustness is central and critical to modern statistical analysis. However, despite the recent advances of deep neural networks (DNNs), many studies have shown that DNNs are vulnerable to adversarial attacks. Making imperceptible changes to an image can cause DNN models to make the wrong classification with high confidence, such as classifying a benign mole as a malignant tumor and a stop sign as a speed limit sign. The trade-off between robustness and standard accuracy is common for DNN models. In this paper, we introduce sensible adversarial learning and demonstrate the synergistic effect between pursuits of standard natural accuracy and robustness. Specifically, we define a sensible adversary which is useful for learning a robust model while keeping high natural accuracy. We theoretically establish that the Bayes classifier is the most robust multi-class classifier with the 0-1 loss under sensible adversarial learning. We propose a novel and efficient algorithm that trains a robust model using implicit loss truncation. We apply sensible adversarial learning for large-scale image classification to a handwritten digital image dataset called MNIST and an object recognition colored image dataset called CIFAR10. We have performed an extensive comparative study to compare our method with other competitive methods. Our experiments empirically demonstrate that our method is not sensitive to its hyperparameter and does not collapse even with a small model capacity while promoting robustness against various attacks and keeping high natural accuracy.

102.PSO-Convolutional Neural Networks with Heterogeneous Learning Rate ⬇️

Convolutional Neural Networks (ConvNets) have been candidly deployed in the scope of computer vision and related fields. Nevertheless, the dynamics of training of these neural networks lie still elusive: it is hard and computationally expensive to train them. A myriad of architectures and training strategies have been proposed to overcome this challenge and address several problems in image processing such as speech, image and action recognition as well as object detection. In this article, we propose a novel Particle Swarm Optimization (PSO) based training for ConvNets. In such framework, the vector of weights of each ConvNet is typically cast as the position of a particle in phase space whereby PSO collaborative dynamics intertwines with Stochastic Gradient Descent (SGD) in order to boost training performance and generalization. Our approach goes as follows: i) [warm-up phase] each ConvNet is trained independently via SGD; ii) [collaborative phase] ConvNets share among themselves their current vector of weights (or particle-position) along with their gradient estimates of the Loss function. Distinct step sizes are coined by distinct ConvNets. By properly blending ConvNets with large (possibly random) step-sizes along with more conservative ones, we propose an algorithm with competitive performance with respect to other PSO-based approaches on Cifar-10 (accuracy of 98.31%). These accuracy levels are obtained by resorting to only four ConvNets -- such results are expected to scale with the number of collaborative ConvNets accordingly. We make our source codes available for download this https URL.

103.Temporally Precise Action Spotting in Soccer Videos Using Dense Detection Anchors ⬇️

We present a model for temporally precise action spotting in videos, which uses a dense set of detection anchors, predicting a detection confidence and corresponding fine-grained temporal displacement for each anchor. We experiment with two trunk architectures, both of which are able to incorporate large temporal contexts while preserving the smaller-scale features required for precise localization: a one-dimensional version of a u-net, and a Transformer encoder (TE). We also suggest best practices for training models of this kind, by applying Sharpness-Aware Minimization (SAM) and mixup data augmentation. We achieve a new state-of-the-art on SoccerNet-v2, the largest soccer video dataset of its kind, with marked improvements in temporal localization. Additionally, our ablations show: the importance of predicting the temporal displacements; the trade-offs between the u-net and TE trunks; and the benefits of training with SAM and mixup.

104.Towards Better Understanding Attribution Methods ⬇️

Deep neural networks are very successful on many vision tasks, but hard to interpret due to their black box nature. To overcome this, various post-hoc attribution methods have been proposed to identify image regions most influential to the models' decisions. Evaluating such methods is challenging since no ground truth attributions exist. We thus propose three novel evaluation schemes to more reliably measure the faithfulness of those methods, to make comparisons between them more fair, and to make visual inspection more systematic. To address faithfulness, we propose a novel evaluation setting (DiFull) in which we carefully control which parts of the input can influence the output in order to distinguish possible from impossible attributions. To address fairness, we note that different methods are applied at different layers, which skews any comparison, and so evaluate all methods on the same layers (ML-Att) and discuss how this impacts their performance on quantitative metrics. For more systematic visualizations, we propose a scheme (AggAtt) to qualitatively evaluate the methods on complete datasets. We use these evaluation schemes to study strengths and shortcomings of some widely used attribution methods. Finally, we propose a post-processing smoothing step that significantly improves the performance of some attribution methods, and discuss its applicability.

105.Combining Contrastive and Supervised Learning for Video Super-Resolution Detection ⬇️

Upscaled video detection is a helpful tool in multimedia forensics, but it is a challenging task that involves various upscaling and compression algorithms. There are many resolution-enhancement methods, including interpolation and deep-learning-based super-resolution, and they leave unique traces. In this work, we propose a new upscaled-resolution-detection method based on learning of visual representations using contrastive and cross-entropy losses. To explain how the method detects videos, we systematically review the major components of our framework - in particular, we show that most data-augmentation approaches hinder the learning of the method. Through extensive experiments on various datasets, we demonstrate that our method effectively detects upscaling even in compressed videos and outperforms the state-of-the-art alternatives. The code and models are publicly available at this https URL

106.Assessing visual acuity in visual prostheses through a virtual-reality system ⬇️

Current visual implants still provide very low resolution and limited field of view, thus limiting visual acuity in implanted patients. Developments of new strategies of artificial vision simulation systems by harnessing new advancements in technologies are of upmost priorities for the development of new visual devices. In this work, we take advantage of virtual-reality software paired with a portable head-mounted display and evaluated the performance of normally sighted participants under simulated prosthetic vision with variable field of view and number of pixels. Our simulated prosthetic vision system allows simple experimentation in order to study the design parameters of future visual prostheses. Ten normally sighted participants volunteered for a visual acuity study. Subjects were required to identify computer-generated Landolt-C gap orientation and different stimulus based on light perception, time-resolution, light location and motion perception commonly used for visual acuity examination in the sighted. Visual acuity scores were recorded across different conditions of number of electrodes and size of field of view. Our results showed that of all conditions tested, a field of view of 20° and 1000 phosphenes of resolution proved the best, with a visual acuity of 1.3 logMAR. Furthermore, performance appears to be correlated with phosphene density, but showing a diminishing return when field of view is less than 20°. The development of new artificial vision simulation systems can be useful to guide the development of new visual devices and the optimization of field of view and resolution to provide a helpful and valuable visual aid to profoundly or totally blind patients.

107.A Dynamic Weighted Tabular Method for Convolutional Neural Networks ⬇️

Traditional Machine Learning (ML) models like Support Vector Machine, Random Forest, and Logistic Regression are generally preferred for classification tasks on tabular datasets. Tabular data consists of rows and columns corresponding to instances and features, respectively. Past studies indicate that traditional classifiers often produce unsatisfactory results in complex tabular datasets. Hence, researchers attempt to use the powerful Convolutional Neural Networks (CNN) for tabular datasets. Recent studies propose several techniques like SuperTML, Conditional GAN (CTGAN), and Tabular Convolution (TAC) for applying Convolutional Neural Networks (CNN) on tabular data. These models outperform the traditional classifiers and substantially improve the performance on tabular data. This study introduces a novel technique, namely, Dynamic Weighted Tabular Method (DWTM), that uses feature weights dynamically based on statistical techniques to apply CNNs on tabular datasets. The method assigns weights dynamically to each feature based on their strength of associativity to the class labels. Each data point is converted into images and fed to a CNN model. The features are allocated image canvas space based on their weights. The DWTM is an improvement on the previously mentioned methods as it dynamically implements the entire experimental setting rather than using the static configuration provided in the previous methods. Furthermore, it uses the novel idea of using feature weights to create image canvas space. In this paper, the DWTM is applied to six benchmarked tabular datasets and it achieves outstanding performance (i.e., average accuracy = 95%) on all of them.

108.Deep Quality Estimation: Creating Surrogate Models for Human Quality Ratings ⬇️

Human ratings are abstract representations of segmentation quality. To approximate human quality ratings on scarce expert data, we train surrogate quality estimation models. We evaluate on a complex multi-class segmentation problem, specifically glioma segmentation following the BraTS annotation protocol. The training data features quality ratings from 15 expert neuroradiologists on a scale ranging from 1 to 6 stars for various computer-generated and manual 3D annotations. Even though the networks operate on 2D images and with scarce training data, we can approximate segmentation quality within a margin of error comparable to human intra-rater reliability. Segmentation quality prediction has broad applications. While an understanding of segmentation quality is imperative for successful clinical translation of automatic segmentation quality algorithms, it can play an essential role in training new segmentation models. Due to the split-second inference times, it can be directly applied within a loss function or as a fully-automatic dataset curation mechanism in a federated learning setting.

109.Contrastive and Non-Contrastive Self-Supervised Learning Recover Global and Local Spectral Embedding Methods ⬇️

Self-Supervised Learning (SSL) surmises that inputs and pairwise positive relationships are enough to learn meaningful representations. Although SSL has recently reached a milestone: outperforming supervised methods in many modalities\dots the theoretical foundations are limited, method-specific, and fail to provide principled design guidelines to practitioners. In this paper, we propose a unifying framework under the helm of spectral manifold learning to address those limitations. Through the course of this study, we will rigorously demonstrate that VICReg, SimCLR, BarlowTwins et al. correspond to eponymous spectral methods such as Laplacian Eigenmaps, Multidimensional Scaling et al.
This unification will then allow us to obtain (i) the closed-form optimal representation for each method, (ii) the closed-form optimal network parameters in the linear regime for each method, (iii) the impact of the pairwise relations used during training on each of those quantities and on downstream task performances, and most importantly, (iv) the first theoretical bridge between contrastive and non-contrastive methods towards global and local spectral embedding methods respectively, hinting at the benefits and limitations of each. For example, (i) if the pairwise relation is aligned with the downstream task, any SSL method can be employed successfully and will recover the supervised method, but in the low data regime, VICReg's invariance hyper-parameter should be high; (ii) if the pairwise relation is misaligned with the downstream task, VICReg with small invariance hyper-parameter should be preferred over SimCLR or BarlowTwins.

110.Orchestra: Unsupervised Federated Learning via Globally Consistent Clustering ⬇️

Federated learning is generally used in tasks where labels are readily available (e.g., next word prediction). Relaxing this constraint requires design of unsupervised learning techniques that can support desirable properties for federated training: robustness to statistical/systems heterogeneity, scalability with number of participants, and communication efficiency. Prior work on this topic has focused on directly extending centralized self-supervised learning techniques, which are not designed to have the properties listed above. To address this situation, we propose Orchestra, a novel unsupervised federated learning technique that exploits the federation's hierarchy to orchestrate a distributed clustering task and enforce a globally consistent partitioning of clients' data into discriminable clusters. We show the algorithmic pipeline in Orchestra guarantees good generalization performance under a linear probe, allowing it to outperform alternative techniques in a broad range of conditions, including variation in heterogeneity, number of clients, participation ratio, and local epochs.

111.Novel Light Field Imaging Device with Enhanced Light Collection for Cold Atom Clouds ⬇️

We present a light field imaging system that captures multiple views of an object with a single shot. The system is designed to maximize the total light collection by accepting a larger solid angle of light than a conventional lens with equivalent depth of field. This is achieved by populating a plane of virtual objects using mirrors and fully utilizing the available field of view and depth of field. Simulation results demonstrate that this design is capable of single-shot tomography of objects of size $\mathcal{O}$(1 mm$^3$), reconstructing the 3-dimensional (3D) distribution and features not accessible from any single view angle in isolation. In particular, for atom clouds used in atom interferometry experiments, the system can reconstruct 3D fringe patterns with size $\mathcal{O}$(100 $\mu$m). We also demonstrate this system with a 3D-printed prototype. The prototype is used to take images of $\mathcal{O}$(1 mm$^{3}$) sized objects, and 3D reconstruction algorithms running on a single-shot image successfully reconstruct $\mathcal{O}$(100 $\mu$m) internal features. The prototype also shows that the system can be built with 3D printing technology and hence can be deployed quickly and cost-effectively in experiments with needs for enhanced light collection or 3D reconstruction. Imaging of cold atom clouds in atom interferometry is a key application of this new type of imaging device where enhanced light collection, high depth of field, and 3D tomographic reconstruction can provide new handles to characterize the atom clouds.

112.Arbitrary Reduction of MRI Slice Spacing Based on Local-Aware Implicit Representation ⬇️

Magnetic resonance (MR) images are often acquired in 2D settings for real clinical applications. The 3D volumes reconstructed by stacking multiple 2D slices have large inter-slice spacing, resulting in lower inter-slice resolution than intra-slice resolution. Super-resolution is a powerful tool to reduce the inter-slice spacing of 3D images to facilitate subsequent visualization and computation tasks. However, most existing works train the super-resolution network at a fixed ratio, which is inconvenient in clinical scenes due to the heterogeneous parameters in MR scanning. In this paper, we propose a single super-resolution network to reduce the inter-slice spacing of MR images at an arbitrarily adjustable ratio. Specifically, we view the input image as a continuous implicit function of coordinates. The intermediate slices of different spacing ratios could be constructed according to the implicit representation up-sampled in the continuous domain. We particularly propose a novel local-aware spatial attention mechanism and long-range residual learning to boost the quality of the output image. The experimental results demonstrate the superiority of our proposed method, even compared to the models trained at a fixed ratio.

113.Collaborative Adversarial Training ⬇️

The vulnerability of deep neural networks (DNNs) to adversarial examples has attracted great attention in the machine learning community. The problem is related to local non-smoothness and steepness of normally obtained loss landscapes. Training augmented with adversarial examples (a.k.a., adversarial training) is considered as an effective remedy. In this paper, we highlight that some collaborative examples, nearly perceptually indistinguishable from both adversarial and benign examples yet show extremely lower prediction loss, can be utilized to enhance adversarial training. A novel method called collaborative adversarial training (CoAT) is thus proposed to achieve new state-of-the-arts.

114.A Coupling Enhancement Algorithm for ZrO2 Ceramic Bearing Ball Surface Defect Detection Based on Cartoon-texture Decomposition Model and Multi-Scale Filtering Method ⬇️

This study aimed to improve the surface defect detection accuracy of ZrO2 ceramic bearing balls. Combined with the noise damage of the image samples, a surface defect detection method for ZrO2 ceramic bearing balls based on cartoon-texture decomposition model was proposed. Building a ZrO2 ceramic bearing ball surface defect detection system. The ZrO2 ceramic bearing ball surface defect image was decomposed by using the Gaussian curvature model and the decomposed image layer was filtered by using Winner filter and wavelet value domain filter. Then they were fused into a clear and undamaged ZrO2 ceramic bearing ball surface defect image and detected. The experimental results show that the image denoising method of ZrO2 ceramic bearing ball surface defect based on cartoon-texture decomposition model can denoise while retaining the image details. The PSNR of image is 34.1 dB, the SSIM is 0.9476, the detection accuracy is 95.8%, and the detection speed of a single defect image is 191ms / img. This method can effectively improve the efficiency and accuracy of ZrO2 ceramic bearing ball surface defect detection.

115.Stability of the scattering transform for deformations with minimal regularity ⬇️

Within the mathematical analysis of deep convolutional neural networks, the wavelet scattering transform introduced by Stéphane Mallat is a unique example of how the ideas of multiscale analysis can be combined with a cascade of modulus nonlinearities to build a nonexpansive, translation invariant signal representation with provable geometric stability properties, namely Lipschitz continuity to the action of small $C^2$ diffeomorphisms - a remarkable result for both theoretical and practical purposes, inherently depending on the choice of the filters and their arrangement into a hierarchical architecture. In this note, we further investigate the intimate relationship between the scattering structure and the regularity of the deformation in the Hölder regularity scale $C^\alpha$, $\alpha >0$. We are able to precisely identify the stability threshold, proving that stability is still achievable for deformations of class $C^{\alpha}$, $\alpha>1$, whereas instability phenomena can occur at lower regularity levels modelled by $C^\alpha$, $0\le \alpha <1$. While the behaviour at the threshold given by Lipschitz (or even $C^1$) regularity remains beyond reach, we are able to prove a stability bound in that case, up to $\varepsilon$ losses.

116.KRNet: Towards Efficient Knowledge Replay ⬇️

The knowledge replay technique has been widely used in many tasks such as continual learning and continuous domain adaptation. The key lies in how to effectively encode the knowledge extracted from previous data and replay them during current training procedure. A simple yet effective model to achieve knowledge replay is autoencoder. However, the number of stored latent codes in autoencoder increases linearly with the scale of data and the trained encoder is redundant for the replaying stage. In this paper, we propose a novel and efficient knowledge recording network (KRNet) which directly maps an arbitrary sample identity number to the corresponding datum. Compared with autoencoder, our KRNet requires significantly ($400\times$) less storage cost for the latent codes and can be trained without the encoder sub-network. Extensive experiments validate the efficiency of KRNet, and as a showcase, it is successfully applied in the task of continual learning.

117.DTU-Net: Learning Topological Similarity for Curvilinear Structure Segmentation ⬇️

Curvilinear structure segmentation plays an important role in many applications. The standard formulation of segmentation as pixel-wise classification often fails to capture these structures due to the small size and low contrast. Some works introduce prior topological information to address this problem with the cost of expensive computations and the need for extra labels. Moreover, prior work primarily focuses on avoiding false splits by encouraging the connection of small gaps. Less attention has been given to avoiding missed splits, namely the incorrect inference of structures that are not visible in the image.
In this paper, we present DTU-Net, a dual-decoder and topology-aware deep neural network consisting of two sequential light-weight U-Nets, namely a texture net, and a topology net. The texture net makes a coarse prediction using image texture information. The topology net learns topological information from the coarse prediction by employing a triplet loss trained to recognize false and missed splits, and provides a topology-aware separation of the foreground and background. The separation is further utilized to correct the coarse prediction. We conducted experiments on a challenging multi-class ultrasound scan segmentation dataset and an open dataset for road extraction. Results show that our model achieves state-of-the-art results in both segmentation accuracy and continuity. Compared to existing methods, our model corrects both false positive and false negative examples more effectively with no need for prior knowledge.

118.Meta-Learning Regrasping Strategies for Physical-Agnostic Objects ⬇️

Grasping inhomogeneous objects, practical use in real-world applications, remains a challenging task due to the unknown physical properties such as mass distribution and coefficient of friction. In this study, we propose a vision-based meta-learning algorithm to learn physical properties in an agnostic way. In particular, we employ Conditional Neural Processes (CNPs) on top of DexNet-2.0. CNPs learn physical embeddings rapidly from a few observations where each observation is composed of i) the cropped depth image, ii) the grasping height between the gripper and estimated grasping point, and iii) the binary grasping result. Our modified conditional DexNet-2.0 (DexNet-CNP) updates the predicted grasping quality iteratively from new observations, which can be executed in an online fashion. We evaluate our method in the Pybullet simulator using various shape primitive objects with different physical parameters. The results show that our model outperforms the original DexNet-2.0 and is able to generalize on unseen objects with different shapes.

119.FedNorm: Modality-Based Normalization in Federated Learning for Multi-Modal Liver Segmentation ⬇️

Given the high incidence and effective treatment options for liver diseases, they are of great socioeconomic importance. One of the most common methods for analyzing CT and MRI images for diagnosis and follow-up treatment is liver segmentation. Recent advances in deep learning have demonstrated encouraging results for automatic liver segmentation. Despite this, their success depends primarily on the availability of an annotated database, which is often not available because of privacy concerns. Federated Learning has been recently proposed as a solution to alleviate these challenges by training a shared global model on distributed clients without access to their local databases. Nevertheless, Federated Learning does not perform well when it is trained on a high degree of heterogeneity of image data due to multi-modal imaging, such as CT and MRI, and multiple scanner types. To this end, we propose Fednorm and its extension \fednormp, two Federated Learning algorithms that use a modality-based normalization technique. Specifically, Fednorm normalizes the features on a client-level, while Fednorm+ employs the modality information of single slices in the feature normalization. Our methods were validated using 428 patients from six publicly available databases and compared to state-of-the-art Federated Learning algorithms and baseline models in heterogeneous settings (multi-institutional, multi-modal data). The experimental results demonstrate that our methods show an overall acceptable performance, achieve Dice per patient scores up to 0.961, consistently outperform locally trained models, and are on par or slightly better than centralized models.

120.An Automated System for Detecting Visual Damages of Wind Turbine Blades ⬇️

Wind energy's ability to compete with fossil fuels on a market level depends on lowering wind's high operational costs. Since damages on wind turbine blades are the leading cause for these operational problems, identifying blade damages is critical. However, recent works in visual identification of blade damages are still experimental and focus on optimizing the traditional machine learning metrics such as IoU. In this paper, we argue that pushing models to production long before achieving the "optimal" model performance can still generate real value for this use case. We discuss the performance of our damage's suggestion model in production and how this system works in coordination with humans as part of a commercialized product and how it can contribute towards lowering wind energy's operational costs.

121.muNet: Evolving Pretrained Deep Neural Networks into Scalable Auto-tuning Multitask Systems ⬇️

Most uses of machine learning today involve training a model from scratch for a particular task, or sometimes starting with a model pretrained on a related task and then fine-tuning on a downstream task. Both approaches offer limited knowledge transfer between different tasks, time-consuming human-driven customization to individual tasks and high computational costs especially when starting from randomly initialized models. We propose a method that uses the layers of a pretrained deep neural network as building blocks to construct an ML system that can jointly solve an arbitrary number of tasks. The resulting system can leverage cross tasks knowledge transfer, while being immune from common drawbacks of multitask approaches such as catastrophic forgetting, gradients interference and negative transfer. We define an evolutionary approach designed to jointly select the prior knowledge relevant for each task, choose the subset of the model parameters to train and dynamically auto-tune its hyperparameters. Furthermore, a novel scale control method is employed to achieve quality/size trade-offs that outperform common fine-tuning techniques. Compared with standard fine-tuning on a benchmark of 10 diverse image classification tasks, the proposed model improves the average accuracy by 2.39% while using 47% less parameters per task.

122.AutoJoin: Efficient Adversarial Training for Robust Maneuvering via Denoising Autoencoder and Joint Learning ⬇️

As a result of increasingly adopted machine learning algorithms and ubiquitous sensors, many 'perception-to-control' systems have been deployed in various settings. For these systems to be trustworthy, we need to improve their robustness with adversarial training being one approach. In this work, we propose a gradient-free adversarial training technique, called AutoJoin. AutoJoin is a very simple yet effective and efficient approach to produce robust models for imaged-based autonomous maneuvering. Compared to other SOTA methods with testing on over 5M perturbed and clean images, AutoJoin achieves significant performance increases up to the 40% range under perturbed datasets while improving on clean performance for almost every dataset tested. In particular, AutoJoin can triple the clean performance improvement compared to the SOTA work by Shen et al. Regarding efficiency, AutoJoin demonstrates strong advantages over other SOTA techniques by saving up to 83% time per training epoch and 90% training data. The core idea of AutoJoin is to use a decoder attachment to the original regression model creating a denoising autoencoder within the architecture. This allows the tasks 'steering' and 'denoising sensor input' to be jointly learnt and enable the two tasks to reinforce each other's performance.

123.Improving AMD diagnosis by the simultaneous identification of associated retinal lesions ⬇️

Age-related Macular Degeneration (AMD) is the predominant cause of blindness in developed countries, specially in elderly people. Moreover, its prevalence is increasing due to the global population ageing. In this scenario, early detection is crucial to avert later vision impairment. Nonetheless, implementing large-scale screening programmes is usually not viable, since the population at-risk is large and the analysis must be performed by expert clinicians. Also, the diagnosis of AMD is considered to be particularly difficult, as it is characterized by many different lesions that, in many cases, resemble those of other macular diseases. To overcome these issues, several works have proposed automatic methods for the detection of AMD in retinography images, the most widely used modality for the screening of the disease. Nowadays, most of these works use Convolutional Neural Networks (CNNs) for the binary classification of images into AMD and non-AMD classes. In this work, we propose a novel approach based on CNNs that simultaneously performs AMD diagnosis and the classification of its potential lesions. This latter secondary task has not yet been addressed in this domain, and provides complementary useful information that improves the diagnosis performance and helps understanding the decision. A CNN model is trained using retinography images with image-level labels for both AMD and lesion presence, which are relatively easy to obtain. The experiments conducted in several public datasets show that the proposed approach improves the detection of AMD, while achieving satisfactory results in the identification of most lesions.

124.Residual Channel Attention Network for Brain Glioma Segmentation ⬇️

A glioma is a malignant brain tumor that seriously affects cognitive functions and lowers patients' life quality. Segmentation of brain glioma is challenging because of interclass ambiguities in tumor regions. Recently, deep learning approaches have achieved outstanding performance in the automatic segmentation of brain glioma. However, existing algorithms fail to exploit channel-wise feature interdependence to select semantic attributes for glioma segmentation. In this study, we implement a novel deep neural network that integrates residual channel attention modules to calibrate intermediate features for glioma segmentation. The proposed channel attention mechanism adaptively weights feature channel-wise to optimize the latent representation of gliomas. We evaluate our method on the established dataset BraTS2017. Experimental results indicate the superiority of our method.

125.Deep Feature Fusion via Graph Convolutional Network for Intracranial Artery Labeling ⬇️

Intracranial arteries are critical blood vessels that supply the brain with oxygenated blood. Intracranial artery labels provide valuable guidance and navigation to numerous clinical applications and disease diagnoses. Various machine learning algorithms have been carried out for automation in the anatomical labeling of cerebral arteries. However, the task remains challenging because of the high complexity and variations of intracranial arteries. This study investigates a novel graph convolutional neural network with deep feature fusion for cerebral artery labeling. We introduce stacked graph convolutions in an encoder-core-decoder architecture, extracting high-level representations from graph nodes and their neighbors. Furthermore, we efficiently aggregate intermediate features from different hierarchies to enhance the proposed model's representation capability and labeling performance. We perform extensive experiments on public datasets, in which the results prove the superiority of our approach over baselines by a clear margin.

126.Producing Histopathology Phantom Images using Generative Adversarial Networks to improve Tumor Detection ⬇️

Advance in medical imaging is an important part in deep learning research. One of the goals of computer vision is development of a holistic, comprehensive model which can identify tumors from histology slides obtained via biopsies. A major problem that stands in the way is lack of data for a few cancer-types. In this paper, we ascertain that data augmentation using GANs can be a viable solution to reduce the unevenness in the distribution of different cancer types in our dataset. Our demonstration showed that a dataset augmented to a 50% increase causes an increase in tumor detection from 80% to 87.5%

127.Scalable and Efficient Training of Large Convolutional Neural Networks with Differential Privacy ⬇️

Large convolutional neural networks (CNN) can be difficult to train in the differentially private (DP) regime, since the optimization algorithms require a computationally expensive operation, known as the per-sample gradient clipping. We propose an efficient and scalable implementation of this clipping on convolutional layers, termed as the mixed ghost clipping, that significantly eases the private training in terms of both time and space complexities, without affecting the accuracy. The improvement in efficiency is rigorously studied through the first complexity analysis for the mixed ghost clipping and existing DP training algorithms.
Extensive experiments on vision classification tasks, with large ResNet, VGG, and Vision Transformers, demonstrate that DP training with mixed ghost clipping adds $1\sim 10%$ memory overhead and $<2\times$ slowdown to the standard non-private training. Specifically, when training VGG19 on CIFAR10, the mixed ghost clipping is $3\times$ faster than state-of-the-art Opacus library with $18\times$ larger maximum batch size. To emphasize the significance of efficient DP training on convolutional layers, we achieve 96.7% accuracy on CIFAR10 and 83.0% on CIFAR100 at $\epsilon=1$ using BEiT, while the previous best results are 94.8% and 67.4%, respectively. We open-source a privacy engine (\url{this https URL}) that implements DP training of CNN with a few lines of code.

128.Transformer based Generative Adversarial Network for Liver Segmentation ⬇️

Automated liver segmentation from radiology scans (CT, MRI) can improve surgery and therapy planning and follow-up assessment in addition to conventional use for diagnosis and prognosis. Although convolutional neural networks (CNNs) have become the standard image segmentation tasks, more recently this has started to change towards Transformers based architectures because Transformers are taking advantage of capturing long range dependence modeling capability in signals, so called attention mechanism. In this study, we propose a new segmentation approach using a hybrid approach combining the Transformer(s) with the Generative Adversarial Network (GAN) approach. The premise behind this choice is that the self-attention mechanism of the Transformers allows the network to aggregate the high dimensional feature and provide global information modeling. This mechanism provides better segmentation performance compared with traditional methods. Furthermore, we encode this generator into the GAN based architecture so that the discriminator network in the GAN can classify the credibility of the generated segmentation masks compared with the real masks coming from human (expert) annotations. This allows us to extract the high dimensional topology information in the mask for biomedical image segmentation and provide more reliable segmentation results. Our model achieved a high dice coefficient of 0.9433, recall of 0.9515, and precision of 0.9376 and outperformed other Transformer based approaches.

129.Equivariant Mesh Attention Networks ⬇️

Equivariance to symmetries has proven to be a powerful inductive bias in deep learning research. Recent works on mesh processing have concentrated on various kinds of natural symmetries, including translations, rotations, scaling, node permutations, and gauge transformations. To date, no existing architecture is equivariant to all of these transformations. Moreover, previous implementations have not always applied these symmetry transformations to the test dataset. This inhibits the ability to determine whether the model attains the claimed equivariance properties. In this paper, we present an attention-based architecture for mesh data that is provably equivariant to all transformations mentioned above. We carry out experiments on the FAUST and TOSCA datasets, and apply the mentioned symmetries to the test set only. Our results confirm that our proposed architecture is equivariant, and therefore robust, to these local/global transformations.

130.A Pilot Study of Relating MYCN-Gene Amplification with Neuroblastoma-Patient CT Scans ⬇️

Neuroblastoma is one of the most common cancers in infants, and the initial diagnosis of this disease is difficult. At present, the MYCN gene amplification (MNA) status is detected by invasive pathological examination of tumor samples. This is time-consuming and may have a hidden impact on children. To handle this problem, we adopt multiple machine learning (ML) algorithms to predict the presence or absence of MYCN gene amplification. The dataset is composed of retrospective CT images of 23 neuroblastoma patients. Different from previous work, we develop the algorithm without manually-segmented primary tumors which is time-consuming and not practical. Instead, we only need the coordinate of the center point and the number of tumor slices given by a subspecialty-trained pediatric radiologist. Specifically, CNN-based method uses pre-trained convolutional neural network, and radiomics-based method extracts radiomics features. Our results show that CNN-based method outperforms the radiomics-based method.

131.Brain Cortical Functional Gradients Predict Cortical Folding Patterns via Attention Mesh Convolution ⬇️

Since gyri and sulci, two basic anatomical building blocks of cortical folding patterns, were suggested to bear different functional roles, a precise mapping from brain function to gyro-sulcal patterns can provide profound insights into both biological and artificial neural networks. However, there lacks a generic theory and effective computational model so far, due to the highly nonlinear relation between them, huge inter-individual variabilities and a sophisticated description of brain function regions/networks distribution as mosaics, such that spatial patterning of them has not been considered. we adopted brain functional gradients derived from resting-state fMRI to embed the "gradual" change of functional connectivity patterns, and developed a novel attention mesh convolution model to predict cortical gyro-sulcal segmentation maps on individual brains. The convolution on mesh considers the spatial organization of functional gradients and folding patterns on a cortical sheet and the newly designed channel attention block enhances the interpretability of the contribution of different functional gradients to cortical folding prediction. Experiments show that the prediction performance via our model outperforms other state-of-the-art models. In addition, we found that the dominant functional gradients contribute less to folding prediction. On the activation maps of the last layer, some well-studied cortical landmarks are found on the borders of, rather than within, the highly activated regions. These results and findings suggest that a specifically designed artificial neural network can improve the precision of the mapping between brain functions and cortical folding patterns, and can provide valuable insight of brain anatomy-function relation for neuroscience.

132.Myocardial Segmentation of Late Gadolinium Enhanced MR Images by Propagation of Contours from Cine MR Images ⬇️

Automatic segmentation of myocardium in Late Gadolinium Enhanced (LGE) Cardiac MR (CMR) images is often difficult due to the intensity heterogeneity resulting from accumulation of contrast agent in infarcted areas. In this paper, we propose an automatic segmentation framework that fully utilizes shared information between corresponding cine and LGE images of a same patient. Given myocardial contours in cine CMR images, the proposed framework achieves accurate segmentation of LGE CMR images in a coarse-to-fine manner. Affine registration is first performed between the corresponding cine and LGE image pair, followed by nonrigid registration, and finally local deformation of myocardial contours driven by forces derived from local features of the LGE image. Experimental results on real patient data with expert outlined ground truth show that the proposed framework can generate accurate and reliable results for myocardial segmentation of LGE CMR images.

133.A Comprehensive 3-D Framework for Automatic Quantification of Late Gadolinium Enhanced Cardiac Magnetic Resonance Images ⬇️

Late gadolinium enhanced (LGE) cardiac magnetic resonance (CMR) can directly visualize nonviable myocardium with hyperenhanced intensities with respect to normal myocardium. For heart attack patients, it is crucial to facilitate the decision of appropriate therapy by analyzing and quantifying their LGE CMR images. To achieve accurate quantification, LGE CMR images need to be processed in two steps: segmentation of the myocardium followed by classification of infarcts within the segmented myocardium. However, automatic segmentation is difficult usually due to the intensity heterogeneity of the myocardium and intensity similarity between the infarcts and blood pool. Besides, the slices of an LGE CMR dataset often suffer from spatial and intensity distortions, causing further difficulties in segmentation and classification. In this paper, we present a comprehensive 3-D framework for automatic quantification of LGE CMR images. In this framework, myocardium is segmented with a novel method that deforms coupled endocardial and epicardial meshes and combines information in both short- and long-axis slices, while infarcts are classified with a graph-cut algorithm incorporating intensity and spatial information. Moreover, both spatial and intensity distortions are effectively corrected with specially designed countermeasures. Experiments with 20 sets of real patient data show visually good segmentation and classification results that are quantitatively in strong agreement with those manually obtained by experts.

134.Unsupervised Sign Language Phoneme Clustering using HamNoSys Notation ⬇️

Traditionally, sign language resources have been collected in controlled settings for specific tasks involving supervised sign classification or linguistic studies accompanied by specific annotation type. To date, very few who explored signing videos found online on social media platforms as well as the use of unsupervised methods applied to such resources. Due to the fact that the field is striving to achieve acceptable model performance on the data that differs from that seen during training calls for more diversity in sign language data, stepping away from the data obtained in controlled laboratory settings. Moreover, since the sign language data collection and annotation carries large overheads, it is desirable to accelerate the annotation process. Considering the aforementioned tendencies, this paper takes the side of harvesting online data in a pursuit for automatically generating and annotating sign language corpora through phoneme clustering.

135.Three-Dimensional Segmentation of the Left Ventricle in Late Gadolinium Enhanced MR Images of Chronic Infarction Combining Long- and Short-Axis Information ⬇️

Automatic segmentation of the left ventricle (LV) in late gadolinium enhanced (LGE) cardiac MR (CMR) images is difficult due to the intensity heterogeneity arising from accumulation of contrast agent in infarcted myocardium. In this paper, we present a comprehensive framework for automatic 3D segmentation of the LV in LGE CMR images. Given myocardial contours in cine images as a priori knowledge, the framework initially propagates the a priori segmentation from cine to LGE images via 2D translational registration. Two meshes representing respectively endocardial and epicardial surfaces are then constructed with the propagated contours. After construction, the two meshes are deformed towards the myocardial edge points detected in both short-axis and long-axis LGE images in a unified 3D coordinate system. Taking into account the intensity characteristics of the LV in LGE images, we propose a novel parametric model of the LV for consistent myocardial edge points detection regardless of pathological status of the myocardium (infarcted or healthy) and of the type of the LGE images (short-axis or long-axis). We have evaluated the proposed framework with 21 sets of real patient and 4 sets of simulated phantom data. Both distance- and region-based performance metrics confirm the observation that the framework can generate accurate and reliable results for myocardial segmentation of LGE images. We have also tested the robustness of the framework with respect to varied a priori segmentation in both practical and simulated settings. Experimental results show that the proposed framework can greatly compensate variations in the given a priori knowledge and consistently produce accurate segmentations.

136.Visualizing CoAtNet Predictions for Aiding Melanoma Detection ⬇️

Melanoma is considered to be the most aggressive form of skin cancer. Due to the similar shape of malignant and benign cancerous lesions, doctors spend considerably more time when diagnosing these findings. At present, the evaluation of malignancy is performed primarily by invasive histological examination of the suspicious lesion. Developing an accurate classifier for early and efficient detection can minimize and monitor the harmful effects of skin cancer and increase patient survival rates. This paper proposes a multi-class classification task using the CoAtNet architecture, a hybrid model that combines the depthwise convolution matrix operation of traditional convolutional neural networks with the strengths of Transformer models and self-attention mechanics to achieve better generalization and capacity. The proposed multi-class classifier achieves an overall precision of 0.901, recall 0.895, and AP 0.923, indicating high performance compared to other state-of-the-art networks.

137.Travel Time, Distance and Costs Optimization for Paratransit Operations using Graph Convolutional Neural Network ⬇️

The provision of paratransit services is one option to meet the transportation needs of Vulnerable Road Users (VRUs). Like any other means of transportation, paratransit has obstacles such as high operational costs and longer trip times. As a result, customers are dissatisfied, and paratransit operators have a low approval rating. Researchers have undertaken various studies over the years to better understand the travel behaviors of paratransit customers and how they are operated. According to the findings of these researches, paratransit operators confront the challenge of determining the optimal route for their trips in order to save travel time. Depending on the nature of the challenge, most research used different optimization techniques to solve these routing problems. As a result, the goal of this study is to use Graph Convolutional Neural Networks (GCNs) to assist paratransit operators in researching various operational scenarios in a strategic setting in order to optimize routing, minimize operating costs and minimize their users' travel time. The study was carried out by using a randomized simulated dataset to help determine the decision to make in terms of fleet composition and capacity under different situations. For the various scenarios investigated, the GCN assisted in determining the minimum optimal gap.

138.Deeper vs Wider: A Revisit of Transformer Configuration ⬇️

Transformer-based models have delivered impressive results on many tasks, particularly vision and language tasks. In many model training situations, conventional configurations are typically adopted. For example, we often set the base model with hidden dimensions (i.e. model width) to be 768 and the number of transformer layers (i.e. model depth) to be 12. In this paper, we revisit these conventional configurations. Through theoretical analysis and experimental evaluation, we show that the masked autoencoder is effective in alleviating the over-smoothing issue in deep transformer training. Based on this finding, we propose Bamboo, an idea of using deeper and narrower transformer configurations, for masked autoencoder training. On ImageNet, with such a simple change in configuration, re-designed model achieves 87.1% top-1 accuracy and outperforms SoTA models like MAE and BEiT. On language tasks, re-designed model outperforms BERT with default setting by 1.1 points on average, on GLUE datasets.

139.Making Video Quality Assessment Models Sensitive to Frame Rate Distortions ⬇️

We consider the problem of capturing distortions arising from changes in frame rate as part of Video Quality Assessment (VQA). Variable frame rate (VFR) videos have become much more common, and streamed videos commonly range from 30 frames per second (fps) up to 120 fps. VFR-VQA offers unique challenges in terms of distortion types as well as in making non-uniform comparisons of reference and distorted videos having different frame rates. The majority of current VQA models require compared videos to be of the same frame rate, but are unable to adequately account for frame rate artifacts. The recently proposed Generalized Entropic Difference (GREED) VQA model succeeds at this task, using natural video statistics models of entropic differences of temporal band-pass coefficients, delivering superior performance on predicting video quality changes arising from frame rate distortions. Here we propose a simple fusion framework, whereby temporal features from GREED are combined with existing VQA models, towards improving model sensitivity towards frame rate distortions. We find through extensive experiments that this feature fusion significantly boosts model performance on both HFR/VFR datasets as well as fixed frame rate (FFR) VQA databases. Our results suggest that employing efficient temporal representations can result much more robust and accurate VQA models when frame rate variations can occur.

140.Enriched Robust Multi-View Kernel Subspace Clustering ⬇️

Subspace clustering is to find underlying low-dimensional subspaces and cluster the data points correctly. In this paper, we propose a novel multi-view subspace clustering method. Most existing methods suffer from two critical issues. First, they usually adopt a two-stage framework and isolate the processes of affinity learning, multi-view information fusion and clustering. Second, they assume the data lies in a linear subspace which may fail in practice as most real-world datasets may have non-linearity structures. To address the above issues, in this paper we propose a novel Enriched Robust Multi-View Kernel Subspace Clustering framework where the consensus affinity matrix is learned from both multi-view data and spectral clustering. Due to the objective and constraints which is difficult to optimize, we propose an iterative optimization method which is easy to implement and can yield closed solution in each step. Extensive experiments have validated the superiority of our method over state-of-the-art clustering methods.

141.Mapping Emulation for Knowledge Distillation ⬇️

This paper formalizes the source-blind knowledge distillation problem that is essential to federated learning. A new geometric perspective is presented to view such a problem as aligning generated distributions between the teacher and student. With its guidance, a new architecture MEKD is proposed to emulate the inverse mapping through generative adversarial training. Unlike mimicking logits and aligning logit distributions, reconstructing the mapping from classifier-logits has a geometric intuition of decreasing empirical distances, and theoretical guarantees using the universal function approximation and optimal mass transportation theories. A new algorithm is also proposed to train the student model that reaches the teacher's performance source-blindly. On various benchmarks, MEKD outperforms existing source-blind KD methods, explainable with ablation studies and visualized results.

142.Masterful: A Training Platform for Computer Vision Models ⬇️

Masterful is a software platform to train deep learning computer vision models. Data and model architecture are inputs to the platform, and the output is a trained model. The platform's primary goal is to maximize a trained model's accuracy, which it achieves through its regularization and semi-supervised learning implementations. The platform's secondary goal is to minimize the amount of manual experimentation typically required to tune training hyperparameters, which it achieves via multiple metalearning algorithms which are custom built to control the platform's regularization and semi-supervised learning implementations. The platform's tertiary goal is to minimize the computing resources required to train a model, which it achieves via another set of metalearning algorithms which are purpose built to control Tensorflow's optimization implementations. The platform builds on top of Tensorflow's data management, architecture, automatic differentiation, and optimization implementations.

143.A SSIM Guided cGAN Architecture For Clinically Driven Generative Image Synthesis of Multiplexed Spatial Proteomics Channels ⬇️

Here we present a structural similarity index measure (SSIM) guided conditional Generative Adversarial Network (cGAN) that generatively performs image-to-image (i2i) synthesis to generate photo-accurate protein channels in multiplexed spatial proteomics images. This approach can be utilized to accurately generate missing spatial proteomics channels that were not included during experimental data collection either at the bench or the clinic. Experimental spatial proteomic data from the Human BioMolecular Atlas Program (HuBMAP) was used to generate spatial representations of missing proteins through a U-Net based image synthesis pipeline. HuBMAP channels were hierarchically clustered by the (SSIM) as a heuristic to obtain the minimal set needed to recapitulate the underlying biology represented by the spatial landscape of proteins. We subsequently prove that our SSIM based architecture allows for scaling of generative image synthesis to slides with up to 100 channels, which is better than current state of the art algorithms which are limited to data with 11 channels. We validate these claims by generating a new experimental spatial proteomics data set from human lung adenocarcinoma tissue sections and show that a model trained on HuBMAP can accurately synthesize channels from our new data set. The ability to recapitulate experimental data from sparsely stained multiplexed histological slides containing spatial proteomic will have tremendous impact on medical diagnostics and drug development, and also raises important questions on the medical ethics of utilizing data produced by generative image synthesis in the clinical setting. The algorithm that we present in this paper will allow researchers and clinicians to save time and costs in proteomics based histological staining while also increasing the amount of data that they can generate through their experiments.

144.Automatic Generation of Synthetic Colonoscopy Videos for Domain Randomization ⬇️

An increasing number of colonoscopic guidance and assistance systems rely on machine learning algorithms which require a large amount of high-quality training data. In order to ensure high performance, the latter has to resemble a substantial portion of possible configurations. This particularly addresses varying anatomy, mucosa appearance and image sensor characteristics which are likely deteriorated by motion blur and inadequate illumination. The limited amount of readily available training data hampers to account for all of these possible configurations which results in reduced generalization capabilities of machine learning models. We propose an exemplary solution for synthesizing colonoscopy videos with substantial appearance and anatomical variations which enables to learn discriminative domain-randomized representations of the interior colon while mimicking real-world settings.

145.Nonlinear motion separation via untrained generator networks with disentangled latent space variables and applications to cardiac MRI ⬇️

In this paper, a nonlinear approach to separate different motion types in video data is proposed. This is particularly relevant in dynamic medical imaging (e.g. PET, MRI), where patient motion poses a significant challenge due to its effects on the image reconstruction as well as for its subsequent interpretation. Here, a new method is proposed where dynamic images are represented as the forward mapping of a sequence of latent variables via a generator neural network. The latent variables are structured so that temporal variations in the data are represented via dynamic latent variables, which are independent of static latent variables characterizing the general structure of the frames. In particular, different kinds of motion are also characterized independently of each other via latent space disentanglement using one-dimensional prior information on all but one of the motion types. This representation allows to freeze any selection of motion types, and to obtain accurate independent representations of other dynamics of interest. Moreover, the proposed algorithm is training-free, i.e., all the network parameters are learned directly from a single video. We illustrate the performance of this method on phantom and real-data MRI examples, where we successfully separate respiratory and cardiac motion.

146.EXPANSE: A Deep Continual / Progressive Learning System for Deep Transfer Learning ⬇️

Deep transfer learning techniques try to tackle the limitations of deep learning, the dependency on extensive training data and the training costs, by reusing obtained knowledge. However, the current DTL techniques suffer from either catastrophic forgetting dilemma (losing the previously obtained knowledge) or overly biased pre-trained models (harder to adapt to target data) in finetuning pre-trained models or freezing a part of the pre-trained model, respectively. Progressive learning, a sub-category of DTL, reduces the effect of the overly biased model in the case of freezing earlier layers by adding a new layer to the end of a frozen pre-trained model. Even though it has been successful in many cases, it cannot yet handle distant source and target data. We propose a new continual/progressive learning approach for deep transfer learning to tackle these limitations. To avoid both catastrophic forgetting and overly biased-model problems, we expand the pre-trained model by expanding pre-trained layers (adding new nodes to each layer) in the model instead of only adding new layers. Hence the method is named EXPANSE. Our experimental results confirm that we can tackle distant source and target data using this technique. At the same time, the final model is still valid on the source data, achieving a promising deep continual learning approach. Moreover, we offer a new way of training deep learning models inspired by the human education system. We termed this two-step training: learning basics first, then adding complexities and uncertainties. The evaluation implies that the two-step training extracts more meaningful features and a finer basin on the error surface since it can achieve better accuracy in comparison to regular training. EXPANSE (model expansion and two-step training) is a systematic continual learning approach applicable to different problems and DL models.

147.Prediction of stent under-expansion in calcified coronary arteries using machine-learning on intravascular optical coherence tomography ⬇️

BACKGROUND Careful evaluation of the risk of stent under-expansions before the intervention will aid treatment planning, including the application of a pre-stent plaque modification strategy.
OBJECTIVES It remains challenging to achieve a proper stent expansion in the presence of severely calcified coronary lesions. Building on our work in deep learning segmentation, we created an automated machine learning approach that uses lesion attributes to predict stent under-expansion from pre-stent images, suggesting the need for plaque modification.
METHODS Pre- and post-stent intravascular optical coherence tomography image data were obtained from 110 coronary lesions. Lumen and calcifications in pre-stent images were segmented using deep learning, and numerous features per lesion were extracted. We analyzed stent expansion along the lesion, enabling frame, segmental, and whole-lesion analyses. We trained regression models to predict the poststent lumen area and then to compute the stent expansion index (SEI). Stents with an SEI < or >/= 80% were classified as "under-expanded" and "well-expanded," respectively.
RESULTS Best performance (root-mean-square-error = 0.04+/-0.02 mm2, r = 0.94+/-0.04, p < 0.0001) was achieved when we used features from both the lumen and calcification to train a Gaussian regression model for a segmental analysis over a segment length of 31 frames. Under-expansion classification results (AUC=0.85+/-0.02) were significantly improved over other approaches.
CONCLUSIONS We used calcifications and lumen features to identify lesions at risk of stent under-expansion. Results suggest that the use of pre-stent images can inform physicians of the need to apply plaque modification approaches.

148.Dual Branch Prior-SegNet: CNN for Interventional CBCT using Planning Scan and Auxiliary Segmentation Loss ⬇️

This paper proposes an extension to the Dual Branch Prior-Net for sparse view interventional CBCT reconstruction incorporating a high quality planning scan. An additional head learns to segment interventional instruments and thus guides the reconstruction task. The prior scans are misaligned by up to +-5deg in-plane during training. Experiments show that the proposed model, Dual Branch Prior-SegNet, significantly outperforms any other evaluated model by >2.8dB PSNR. It also stays robust wrt. rotations of up to +-5.5deg.