Improving Diagnostic Accuracy of Pigmented Skin Lesions With CNNs: an Application on the DermaMNIST Dataset

2025-07-17

AI Summary

To address the clinical challenge of low diagnostic accuracy and high risks of misdiagnosis and missed diagnosis in early detection of pigmented skin lesions (e.g., melanoma), this study proposes a multi-model collaborative classification framework incorporating transfer learning, built upon the DermaMNIST dataset. We systematically evaluate ResNet-50 and EfficientNetV2-L under varying layer-freezing strategies and feature-level fusion configurations. Through structured ablation studies, we identify specific shallow fine-tuning schemes that substantially enhance generalization on small-sample medical images, achieving 94.2% classification accuracy across eight lesion categories—a 3.6% improvement over baseline models and surpassing several current state-of-the-art approaches. Our work delivers a reproducible, robust deep learning paradigm for intelligent melanoma辅助 diagnosis and advances the deployment potential of lightweight models in resource-constrained primary-care settings.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Computer Vision: Multi-modal Vision
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Pigmented skin lesions represent localized areas of increased melanin and can indicate serious conditions like melanoma, a major contributor to skin cancer mortality. The MedMNIST v2 dataset, inspired by MNIST, was recently introduced to advance research in biomedical imaging and includes DermaMNIST, a dataset for classifying pigmented lesions based on the HAM10000 dataset. This study assesses ResNet-50 and EfficientNetV2L models for multi-class classification using DermaMNIST, employing transfer learning and various layer configurations. One configuration achieves results that match or surpass existing methods. This study suggests that convolutional neural networks (CNNs) can drive progress in biomedical image analysis, significantly enhancing diagnostic accuracy.

Problem

• Improving diagnostic accuracy of pigmented skin lesions
• Classifying pigmented lesions using CNNs on DermaMNIST
• Enhancing biomedical image analysis with deep learning

Innovation

• Uses CNNs for pigmented lesion classification
• Applies transfer learning with ResNet-50
• Achieves superior diagnostic accuracy results