Explainable AI: Comparative Analysis of Normal and Dilated ResNet Models for Fundus Disease Classification

TL;DR

The paper investigates whether dilated ResNet architectures improve multiclass fundus disease classification over standard ResNet variants when trained on the ODIR-5K dataset. By inserting dilation in the higher blocks of ResNet models (five depths) and using the SCCE loss, the authors show that dilation enhances F1 scores and accuracy, with the deepest networks achieving the largest gains. They also apply explainable AI techniques (LIME, RISE, Grad-CAM) and activation-map analysis to verify that the model's focus aligns with clinically relevant retinal features, such as the optic disc, macula, and retinal vessels. Overall, the results support using dilated convolution in high-level ResNet blocks for improved diagnostic performance, while XAI analyses provide transparency that can facilitate clinical adoption.

Abstract

This paper presents dilated Residual Network (ResNet) models for disease classification from retinal fundus images. Dilated convolution filters are used to replace normal convolution filters in the higher layers of the ResNet model (dilated ResNet) in order to improve the receptive field compared to the normal ResNet model for disease classification. This study introduces computer-assisted diagnostic tools that employ deep learning, enhanced with explainable AI techniques. These techniques aim to make the tool's decision-making process transparent, thereby enabling medical professionals to understand and trust the AI's diagnostic decision. They are particularly relevant in today's healthcare landscape, where there is a growing demand for transparency in AI applications to ensure their reliability and ethical use. The dilated ResNet is used as a replacement for the normal ResNet to enhance the classification accuracy of retinal eye diseases and reduce the required computing time. The dataset used in this work is the Ocular Disease Intelligent Recognition (ODIR) dataset which is a structured ophthalmic database with eight classes covering most of the common retinal eye diseases. The evaluation metrics used in this work include precision, recall, accuracy, and F1 score. In this work, a comparative study has been made between normal ResNet models and dilated ResNet models on five variants namely ResNet-18, ResNet-34, ResNet-50, ResNet-101, and ResNet-152. The dilated ResNet model shows promising results as compared to normal ResNet with an average F1 score of 0.71, 0.70, 0.69, 0.67, and 0.70 respectively for the above respective variants in ODIR multiclass disease classification.

Figures (21)

• Figure 1: Source: Global Health Matters, 2012 GHM2012
• Figure 2: Retinal fundus image
• Figure 3: Normal convolution vs Dilated convolution in 2D
• Figure 4: Flow diagram of the proposed work
• Figure 5: ResNet-18 model (a) Normal convolution (b) Dilated convolution