Visual Bias and Interpretability in Deep Learning for Dermatological Image Analysis

TL;DR

This paper evaluates various deep learning models and image pre-processing techniques for skin disease classification, demonstrating that transformer-based models like DinoV2 with RGB pre-processing yield high accuracy and interpretability in dermatological image analysis.

Contribution

It systematically benchmarks pre-processing methods and deep learning architectures, highlighting the effectiveness of transformer models and RGB pre-processing for skin disease classification.

Findings

DinoV2 with RGB achieves up to 93% accuracy.

Grad-CAM visualizations improve interpretability.

Pre-processing significantly impacts model performance.

Abstract

Accurate skin disease classification is a critical yet challenging task due to high inter-class similarity, intra-class variability, and complex lesion textures. While deep learning-based computer-aided diagnosis (CAD) systems have shown promise in automating dermatological assessments, their performance is highly dependent on image pre-processing and model architecture. This study proposes a deep learning framework for multi-class skin disease classification, systematically evaluating three image pre-processing techniques: standard RGB, CMY color space transformation, and Contrast Limited Adaptive Histogram Equalization (CLAHE). We benchmark the performance of pre-trained convolutional neural networks (DenseNet201, Efficient-NetB5) and transformer-based models (ViT, Swin Transformer, DinoV2 Large) using accuracy and F1-score as evaluation metrics. Results show that DinoV2 with RGB pre-processing achieves the highest accuracy (up to 93%) and F1-scores across all variants. Grad-CAM visualizations applied to RGB inputs further reveal precise lesion localization, enhancing interpretability. These findings underscore the importance of effective pre-processing and model choice in building robust and explainable CAD systems for dermatology.