Leveraging Spatial and Semantic Feature Extraction for Skin Cancer Diagnosis with Capsule Networks and Graph Neural Networks

Overview

This repository contains the implementation and supplementary materials for the paper titled "Leveraging Spatial and Semantic Feature Extraction for Skin Cancer Diagnosis with Capsule Networks and Graph Neural Networks." To build this model, we adopted Han, et. al. Vision GNN and incorporated capsule networks to analyze its performance on skin lesion classification.

Abstract

Skin lesion classification presents challenges due to intricate spatial and semantic features and imbalanced datasets. This paper introduces a hybrid model combining Graph Neural Networks (GNNs) and Capsule Networks to enhance classification performance. Our model, applied to the MNIST:HAM10000 dataset, achieved superior accuracy compared to benchmarks, demonstrating significant improvements in diagnosing skin cancer.

Authors

• K. P. Santoso
• R. V. H. Ginardi
• R. A. Sastrowardoyo
• F. A. Madany

Contributions

• Integration of GNNs and Capsule Networks
• Application to MNIST:HAM10000 skin lesion dataset
• Achieved accuracy: 95.52%

Results

Our model outperformed established benchmarks:

• GoogLeNet: 83.94%
• InceptionV3: 86.82%
• MobileNet V3: 89.87%
• EfficientNet-B7: 92.07%
• ResNet18: 92.22%
• ResNet34: 91.90%
• ViT-Base: 73.70%
• IRv2-SA: 93.47%

Installation

Clone the repository and install the required packages:

git clone https://github.com/echelon2718/ViG-GNN-HAM10000.git
cd skin-cancer-diagnosis
pip install -r requirements.txt