MelanomaNet: Explainable Deep Learning for Skin Lesion Classification

TL;DR

MelanomaNet is an explainable deep learning system for skin lesion classification that combines multiple interpretability methods, achieving high accuracy while providing clinically meaningful explanations and uncertainty estimates to support clinical adoption.

Contribution

The paper introduces MelanomaNet, integrating four interpretability mechanisms with a deep learning model for skin lesion classification, enhancing transparency and trust in clinical settings.

Findings

Achieves 85.61% accuracy on ISIC 2019 dataset

Provides clinically aligned explanations via GradCAM++ and concept vectors

Includes uncertainty quantification to flag unreliable predictions

Abstract

Automated skin lesion classification using deep learning has shown remarkable accuracy, yet clinical adoption remains limited due to the "black box" nature of these models. We present MelanomaNet, an explainable deep learning system for multi-class skin lesion classification that addresses this gap through four complementary interpretability mechanisms. Our approach combines an EfficientNet V2 backbone with GradCAM++ attention visualization, automated ABCDE clinical criterion extraction, Fast Concept Activation Vectors (FastCAV) for concept-based explanations, and Monte Carlo Dropout uncertainty quantification. We evaluate our system on the ISIC 2019 dataset containing 25,331 dermoscopic images across 9 diagnostic categories. Our model achieves 85.61% accuracy with a weighted F1 score of 0.8564, while providing clinically meaningful explanations that align model attention with established dermatological assessment criteria. The uncertainty quantification module decomposes prediction confidence into epistemic and aleatoric components, enabling automatic flagging of unreliable predictions for clinical review. Our results demonstrate that high classification performance can be achieved alongside comprehensive interpretability, potentially facilitating greater trust and adoption in clinical dermatology workflows. The source code is available at https://github.com/suxrobgm/explainable-melanoma