Exploring the Challenge and Value of Deep Learning in Automated Skin Disease Diagnosis

TL;DR

The paper surveys 91 studies (2020–2025) on deep learning for automated skin disease diagnosis, identifying four core challenges: complex features, image noise, intra-class variation, inter-class similarity, and data imbalance. It analyzes how data augmentation, hybrid CNN-Transformer architectures, and multimodal feature fusion address these challenges, with emphasis on transfer learning, attention mechanisms, and optimized loss functions to improve balance, interpretability, and clinical applicability. Using a PRISMA-based methodology, it analyzes benchmark datasets (ISIC, HAM10000, PH2, Derm7pt, SkinCon) and identifies a trend toward clinical translation aided by advanced augmentation, robust feature integration, and explainable AI. The review highlights forward-looking directions like GAN/diffusion-based augmentation, adaptive hybrid models, and cross-modal data synthesis to enhance robustness and adoption in real-world dermatology.

Abstract

Skin cancer is one of the most prevalent and deadly forms of cancer worldwide, which highlights the critical importance of early detection and diagnosis in improving patient outcomes. Deep learning (DL) has shown significant promise in enhancing the accuracy and efficiency of automated skin disease diagnosis, particularly in detecting and evaluating skin lesions and classification. However, there are still several challenges for DL-based skin cancer diagnosis, including complex features, image noise, intra-class variation, inter-class similarity, and data imbalance. By synthesizing recent research, this review discusses innovative approaches to cope with these challenges, such as data augmentation, hybrid models, and feature fusion, etc. Furthermore, the review highlights the integration of DL models into clinical workflows, offering insights into the potential of deep learning to revolutionize skin disease diagnosis and improve clinical decision-making. This article follows a comprehensive methodology based on the PRISMA framework and emphasizes the need for continued advancements to fully unlock the transformative potential of DL in dermatological care.

Figures (8)

• Figure 1: Challenges of current deep learning studies.
• Figure 2: The PRISMA flow of this literature.
• Figure 3: Publication Trend of research from 2020 to June 2025.
• Figure 4: Challenges of current deep learning studies
• Figure 5: Global distribution of first authors’ affiliations.