Awesome Scene Understanding

A list of papers for scene understanding.

Workshops and Tutorials

• Holistic Structures for 3D Vision (ICCV'21) [Webpage]

• Holistic Scene Structures for 3D Vision (ECCV'20) [Webpage] [Challenge]

• Holistic 3D Reconstruction: Learning to Reconstruct Holistic 3D Structures from Sensorial Data (ICCV'19) [Webpage] [Resources]

Survey

• State-of-the-art in Automatic 3D Reconstruction of Structured Indoor Environments (Computer Graphics Forum'20) [Paper]

• Indoor Scene Understanding in 2.5/3D for Autonomous Agents: A Survey (IEEE Access'19) [Paper]

• RGBD Datasets: Past, Present and Future (CVPR Workshop'16) [Project] [Paper]

Dataset

Realistic Dataset

• HoliCity: A City-Scale Data Platform for Learning Holistic 3D Structures (CoRR'20) [Project] [Paper] [Code]

• OASIS: A Large-Scale Dataset for Single Image 3D in the Wild (CVPR'20) [Project] [Paper]

• 3D Scene Graph: A Structure for Unified Semantics, 3D Space, and Camera (ICCV'19) [Project] [Paper]

• The Replica Dataset: A Digital Replica of Indoor Spaces (CoRR'19) [Paper] [Code]

• Matterport3D: Learning from RGB-D Data in Indoor Environments (3DV'17) [Project] [Paper] [Code]

• [2D-3D-S] Joint 2D-3D-Semantic Data for Indoor Scene Understanding (CoRR'17) [Project] [Paper]

• ScanNet: Richly-annotated 3D Reconstructions of Indoor Scenes (CVPR'17) [Project] [Paper]

• SceneNN: a Scene Meshes Dataset with aNNotations (3DV'16) [Project] [Paper]

• SUN RGB-D: A RGB-D Scene Understanding Benchmark Suite (CVPR'15) [Project] [Paper]

• SUN3D: A Database of Big Spaces Reconstructed using SfM and Object Labels (ICCV'13) [Project] [Paper]

• [NYUv2] Indoor Segmentation and Support Inference from RGBD Images (ECCV'12) [Project] [Paper]

Synthetic Dataset

• MINERVAS: Massive INterior EnviRonments VirtuAl Synthesis (CoRR'21) [Project] [Paper]

• 3D-FRONT: 3D Furnished Rooms with layOuts and semaNTics (ICCV'21) [Project] [Paper] [Code] [Rendering Tool]

• Hypersim: A Photorealistic Synthetic Dataset for Holistic Indoor Scene Understanding (ICCV'21) [Project] [Paper] [Code]

• OpenRooms: An End-to-End Open Framework for Photorealistic Indoor Scene Datasets (CVPR'21) [Project] [Paper]

• Structured3D: A Large Photo-realistic Dataset for Structured 3D Modeling (ECCV'20) [Project] [Paper] [Code]

• InteriorNet: Mega-scale Multi-sensor Photo-realistic Indoor Scenes Dataset (BMVC'18) [Project] [Paper]

• SceneNet RGB-D: Can 5M Synthetic Images Beat Generic ImageNet Pre-training on Indoor Segmentation? (ICCV'17) [Project] [Paper]

• [SUNCG] Semantic Scene Completion from a Single Depth Image (CVPR'17) [Paper]

• SceneNet: Understanding Real World Indoor Scenes With Synthetic Data (CVPR'16) [Project] [Paper]

• The SYNTHIA Dataset: A Large Collection of Synthetic Images for Semantic Segmentation of Urban Scenes (CVPR'16) [Project] [Paper]

Holistic Scene Understanding

Perspective Image

• Holistic 3D Scene Understanding from a Single Image with Implicit Representation (CVPR'21) [Paper]

• Total3DUnderstanding: Joint Layout, Object Pose and Mesh Reconstruction for Indoor Scenes from a Single Image (CVPR'20) [Paper] [Code]

• 3D Object Detection from a Single RGB Image via Perspective Points (NIPS'19) [Paper]

• Hoilistc++ Scene Understanding: Single-view 3D Holistic Scene Parsing and Human Pose Estimation with Human-Object Interaction and Physical Commonsense (ICCV'19) [Project] [Paper] [Code]

• Complete 3D Scene Parsing from an RGBD Image (IJCV'18) [Paper]

• Cooperative Holistic Scene Understanding: Unifying 3D Object, Layout, and Camera Pose Estimation (NeurIPS'18) [Project] [Paper] [Code]

• Holistic 3D Scene Parsing and Reconstruction from a Single RGB Image (ECCV'18) [Project] [Paper] [Code]

• Factoring Shape, Pose, and Layout from the 2D Image of a 3D Scene (CVPR'18) [Project] [Paper] [Code]

• Im2CAD (CVPR'18) [Project] [Paper]

• DeepContext: Context-Encoding Neural Pathways for 3D Holistic Scene Understanding (ICCV'17) [Project] [Paper]

• Emptying, Refurnishing, and Relighting Indoor Spaces (SIGGRAPH Asia'16) [Project] [Paper]

• Scene Parsing by Integrating Function, Geometry and Appearance Models (CVPR'13) [Project] [Paper]

• Understanding Indoor Scenes using 3D Geometric Phrases (CVPR'13) [Paper]

• Recovering Free Space of Indoor Scenes from a Single Image (CVPR'12) [Paper]

• Efficient Exact Inference for 3D Indoor Scene Understanding (ECCV'12) [Paper]

• Efficient Structured Prediction for 3D Indoor Scene Understanding (CVPR'12) [Paper]

• Estimating Spatial Layout of Rooms using Volumetric Reasoning about Objects and Surfaces (NeurIPS'10) [Paper]

• Thinking Inside the Box: Using Appearance Models and Context Based on Room Geometry (ECCV'10) [Paper]

Panoramic Image

• DeepPanoContext: Panoramic 3D Scene Understanding with Holistic Scene Context Graph and Relation-based Optimization (ICCV'21) [Paper] [Code]

• Automatic 3D Indoor Scene Modeling from Single Panorama (CVPR'18) [Paper]

• Pano2CAD: Room Layout From A Single Panorama Image (WACV'17) [Paper]

• PanoContext: A Whole-room 3D Context Model for Panoramic Scene Understanding (ECCV'14) [Project] [Paper]

Room Layout Estimation

Perspective Image

(AW: Atlanta-world, SS: single-floor and single-ceiling, PP: Piece-wise Planarity.)

Dataset	Modality	#Frames	Prior	Source
Hedau (ICCV'09)	RGB	366	Cuboid	-
NYUv2 303 (ICCV'13)	RGB-D	303	Cuboid	NYUv2
LSUN Room Layout (2016)	RGB	5394	Cuboid	SUN
SUN RGB-D (CVPR'15)	RGB-D	10335	AW+SS	NYUv2, Berkeley B3DO, and SUN3D
ScanNet-Layout (ECCV'20)	RGB-D	293	PP	ScanNet
Matterport3D-Layout (ECCV'20)	RGB-D	7360	PP	Matterport
Structured3D (ECCV'20)	RGB-D	82027	AW+SS	Structured3D

• Learning to Reconstruct 3D Non-Cuboid Room Layout from a Single RGB Image (WACV'22) [Paper] [Code]

• RoomStructNet: Learning to Rank Non-Cuboidal Room Layouts From Single View (CoRR'21) [Paper]

• GeoLayout: Geometry Driven Room Layout Estimation Based on Depth Maps of Planes (ECCV'20) [Paper] [Matterport3D Layout Dataset]

• Structural Deep Metric Learning for Room Layout Estimation (ECCV'20) [Paper]

• General 3D Room Layout from a Single View by Render-and-Compare (ECCV'20) [Project] [Paper] [ScanNet-Layout Dataset] [Code]

• Smart Hypothesis Generation for Efficient and Robust Room Layout Estimation (WACV'20) [Paper]

• Flat2Layout: Flat Representation for Estimating Layout of General Room Types (CoRR'19) [Paper]

• Thinking Outside the Box: Generation of Unconstrained 3D Room Layouts (ACCV'18) [Paper]

• RoomNet: End-to-End Room Layout Estimation (ICCV'17) [Paper]

• Physics Inspired Optimization on Semantic Transfer Features: An Alternative Method for Room Layout Estimation (CVPR'17) [Project] [Paper]

• A Coarse-to-Fine Indoor Layout Estimation (CFILE) Method (ACCV'16) [Paper]

• DeLay: Robust Spatial Layout Estimation for Cluttered Indoor Scenes (CVPR'16) [Paper]

• Learning Informative Edge Maps for Indoor Scene Layout Prediction (ICCV'15) [Homepage] [Paper]

• Rent3D: Floor-Plan Priors for Monocular Layout Estimation (CVPR'15) [Project] [Paper]

• Box In the Box: Joint 3D Layout and Object Reasoning from Single Images (CVPR'13) [Paper]

• Estimating the 3D Layout of Indoor Scenes and its Clutter from Depth Sensors (ICCV'13) [Project] [Paper]

• Recovering the Spatial Layout of Cluttered Rooms (ICCV'09) [Paper]

Panoramic Image

(MW: Manhattan world, AW: Atlanta world, SS: single-floor and single-ceiling.)

Dataset	Modality	#Frames	Prior	Source
PanoContext (ECCV'14)	RGB	500	Cuboid	SUN360
2D-3D-S (CVPR'18)	RGB-D	571	Cuboid	2D-3D-S
MatterportLayout (2020)	RGB-D	2295	MW+SS	Matterport
LayoutMP3D (2020)	RGB-D	2505	MW+SS	Matterport
Structured3D (ECCV'20)	RGB-D	196515	AW+SS	Structured3D
ZInD (CVPR'21)	RGB	71474	AW+SS	ZinD

• Transferable End-to-end Room Layout Estimation via Implicit Encoding (CoRR'21) [Paper] [Project]

• Zillow Indoor Dataset: Annotated Floor Plans With 360˚ Panoramas and 3D Room Layouts (CVPR'21) [Paper] [Code]

• OmniLayout: Room Layout Reconstruction from Indoor Spherical Panoramas (CVPR Workshop'21) [Paper] [Code]

• LED²-Net: Monocular 360˚ Layout Estimation via Differentiable Depth Rendering (CVPR'21) [Project] [Paper] [Code]

• SSLayout360: Semi-Supervised Indoor Layout Estimation from 360 Panorama (CVPR'21) [Paper]

• Single-Shot Cuboids: Geodesics-based End-to-end Manhattan Aligned Layout Estimation from Spherical Panoramas (CoRR'21) [Project] [Paper] [Code]

• Manhattan Room Layout Reconstruction from a Single 360 image: A Comparative Study of State-of-the-art Methods (IJCV'21) [Paper] [Code] [MatterportLayout Dataset]

• Training and Post Processing 3D Room Layout Beyond the Manhattan World Assumption (ECCV Workshop'20) [Paper]

• Joint 3D Layout and Depth Prediction from a Single Indoor Panorama Image (ECCV'20) [Paper]

• AtlantaNet: Inferring the 3D Indoor Layout from a Single 360 Image Beyond the Manhattan World Assumption (ECCV'20) [Project] [Paper] [Code]

• Corners for Layout: End-to-End Layout Recovery from 360 Images (ICRA'19) [Project] [Paper] [Code]

• DuLa-Net: A Dual-Projection Network for Estimating Room Layouts from a Single RGB Panorama (CVPR'19) [Project] [Paper]

• HorizonNet: Learning Room Layout with 1D Representation and Pano Stretch Data Augmentation (CVPR'19) [Paper] [Code]

• Layouts from Panoramic Images with Geometry and Deep Learning (IROS'18) [Paper] [Code]

• LayoutNet: Reconstructing the 3D Room Layout from a Single RGB Image (CVPR'18) [Paper] [Code]

• Efficient 3D Room Shape Recovery From a Single Panorama (CVPR'16) [Project] [Paper] [Code]

Floorplan

Point Cloud

• MonteFloor: Extending MCTS for Reconstructing Accurate Large-Scale Floor Plans (ICCV'21) [Paper]

• Scan2Plan: Efficient Floorplan Generation from 3D Scans of Indoor Scenes (CoRR'20) [Paper]

• Floor-SP: Inverse CAD for Floorplans by Sequential Room-wise Shortest Path (ICCV'19) [Project] [Paper] [Code]

• FloorNet: A unified framework for floorplan reconstruction from 3D scans (ECCV'18) [Project] [Paper] [Code]

Multi-view

• Extreme Structure From Motion for Indoor Panoramas Without Visual Overlaps (ICCV'21) [Paper] [Code]

• Floorplan-Jigsaw: Jointly Estimating Scene Layout and Aligning Partial Scans (ICCV'19) [Project] [Paper]

• DeepPerimeter: Indoor Boundary Estimation from Posed Monocular Sequences (CoRR'19) [Paper]

Image

• Residential floor plan recognition and reconstruction (CVPR'21) [Paper]

• Versailles-FP dataset: Wall Detection in Ancient Floor Plans (CoRR'21) [Paper]

• HouseExpo: A Large-scale 2D Indoor Layout Dataset for Learning-based Algorithms on Mobile Robots (IROS'20) [Paper] [Code]

• Deep Floor Plan Recognition using a Multi-task Network with Room-boundary-Guided Attention (ICCV'19) [Project] [Paper]

• CubiCasa5K: A Dataset and an Improved Multi-Task Model for Floorplan Image Analysis (CoRR'19) [Paper] [Code]

• Raster-to-Vector: Revisiting Floorplan Transformation (ICCV'17) [Project] [Paper] [Code]

Primitive Detection

Junction

• Manhattan Junction Catalogue for Spatial Reasoning of Indoor Scenes (CVPR'13) [Paper]

Line Segment and Wireframe

Please refer to Wireframe to see more comprehensive review.

• Towards Real-time and Light-weight Line Segment Detection (AAAI'22) [Paper] [Code]

• Hole-robust Wireframe Detection (WACV'22) [Paper]

• Fully Convolutional Line Parsing (CoRR'21) [Paper] [Code]

• ELSD: Efficient Line Segment Detector and Descriptor (ICCV'21) [Paper]

• SOLD²: Self-supervised Occlusion-aware Line Description and Detection (CVPR'21) [Paper] [Code]

• Line Segment Detection Using Transformers without Edges (CVPR'21) [Paper]

• PlueckerNet: Learn to Register 3D Line Reconstructions (CVPR'20) [Paper] [Code]

• LGNN: A Context-aware Line Segment Detector (ACM MM'20) [Paper]

• TP-LSD: Tri-Points Based Line Segment Detector (ECCV'20) [Paper]

• Deep Hough-Transform Line Priors (ECCV'20) [Paper] [Code]

• Deep Hough Transform for Semantic Line Detection (ECCV'20) [Paper] [Code]

• Holistically-Attracted Wireframe Parsing (CVPR'20) [Paper] [Code]

• Learning to Reconstruct 3D Manhattan Wireframes from a Single Image (ICCV'19) [Paper] [Code]

• End-to-End Wireframe Parsing (ICCV'19) [Paper] [Code]

• PPGNet: Learning Point-Pair Graph for Line Segment Detection (CVPR'19) [Paper] [Code]

• Learning Attraction Field Representation for Robust Line Segment Detection (CVPR'19) [Paper] [Code]

• Novel Single View Constraints for Manhattan 3D Line Reconstruction (3DV'18) [Paper]

• Learning to Parse Wireframes in Images of Man-Made Environments (CVPR'18) [Paper] [Code]

• A Novel Linelet-Based Representation for Line Segment Detection (TPAMI'18) [Paper]

• MCMLSD: A Dynamic Programming Approach to Line Segment Detection (CVPR'17) [Paper]

• Lifting 3D Manhattan Lines from a Single Image (ICCV'15) [Paper]

• LSD: A Fast Line Segment Detector with a False Detection Control (TPAMI'10) [Paper]

Plane

• PlaneRecNet: Multi-Task Learning with Cross-Task Consistency for Piece-Wise Plane Detection and Reconstruction from a Single RGB Image (BMVC'21) [Paper] [Code]

• PlaneTR: Structure-Guided Transformers for 3D Plane Recovery (ICCV'21) [Paper] [Code]

• Planar Surface Reconstruction From Sparse Views (ICCV'21) [Project] [Paper] [Code]

• Indoor Panorama Planar 3D Reconstruction via Divide and Conquer (CVPR'21) [Paper] [Code]

• Learning Pairwise Inter-Plane Relations for Piecewise Planar Reconstruction (ECCV'20) [Paper] [Code]

• Peek-a-Boo: Occlusion Reasoning in Indoor Scenes with Plane Representations (CVPR'20) [Project] [Paper]

• Single-Image Piece-wise Planar 3D Reconstruction via Associative Embedding (CVPR'19) [Paper] [Code]

• PlaneRCNN: 3D Plane Detection and Reconstruction from a Single Image (CVPR'19) [Project] [Paper] [Code]

• Recovering 3D Planes from a Single Image via Convolutional Neural Networks (ECCV'18) [Paper] [Code]

• PlaneNet: Piece-wise Planar Reconstruction from a Single RGB Image (CVPR'18) [Project] [Paper] [Code]

Cuboid

• Deep Cuboid Detection: Beyond 2D Bounding Boxes (CoRR'16) [Paper]

• A Linear Approach to Matching Cuboids in RGBD Images (CVPR'13) [Project] [Paper]

• Localizing 3D Cuboids in Single-view Images (NIPS'12) [Paper]

Others

• Bottom-Up/Top-Down Image Parsing with Attribute Grammar (TPAMI'09) [Paper]

• Detection and Matching of Rectilinear Structures (CVPR'08) [Paper]