MDSkin-Net: Multi-Task Skin Lesion Analysis Driven by Pattern Analysis Priors and Spatial Alignment Regularization

2026-09-25

AI Summary

This study addresses the limitations of existing dermoscopic multi-task frameworks, which typically lack clinical priors and exhibit loose coupling between segmentation and classification. To overcome these issues, we propose a hybrid CNN-Transformer architecture that integrates pattern analysis priors with spatial alignment regularization. The core innovations include a pattern analysis-guided attention module that translates macroscopic clinical rules into micro-level clue-based priors, as well as multi-scale spatial alignment regularization, enhanced channel attention, and biased asymmetric attention mechanisms to achieve deep synergy between segmentation and classification. Evaluated on the PH2 dataset, the proposed method attains a segmentation Dice similarity coefficient of 92.38% and a melanoma classification AUC of 97.84%, while demonstrating superior cross-domain zero-shot generalization capabilities.

Abstract

Reliable skin lesion segmentation and classification are central to dermoscopic computer-aided diagnosis. Existing multi-task frameworks couple the two tasks architecturally without clinical knowledge, while knowledge-injecting approaches rely on the macroscopic ABCD rule, which was not designed for dermoscopy. Dermoscopic diagnosis is grounded in Pattern Analysis, a microscopic framework structured around dermoscopic features. We propose MDSkin-Net, which incorporates cue-level Pattern Analysis priors into a hybrid CNN-Transformer architecture. At its core is a Pattern Analysis-Guided Attention Module (PAGAM) comprising three priors motivated by distinct dermoscopic cues: an improved Efficient Channel Attention (iECA), a Multi-Scale Spatial Attention (MSSA), and a Biased Asymmetry Attention (BAA). We further introduce a multi-scale spatial alignment regularization (MSAR) that uses the segmentation ground-truth mask as hierarchical soft supervision, confining the classification head to lesion-localized evidence and coupling both task pathways through a shared spatial prior. Trained exclusively on the ISIC 2017 training split without external dermoscopy data, the MDSkin-Net ensemble transfers robustly under zero-shot evaluation, reaching a Dice Similarity Coefficient (DSC) of 92.38% and a melanoma AUC of 97.84%on PH2, and a DSC of 88.92% on the ISIC 2018 Task 1 test set. On the in-domain ISIC 2017 benchmark, the ensemble attains a mean Area Under the Curve (AUC) of 91.60% across the two classification tasks (melanoma and seborrheic keratosis vs. rest), and a DSC of 84.72% for segmentation. Classification remains competitive with baselines; in-domain segmentation trails single-task specialists, yet the proposed priors and alignment regularization yield representations that generalize consistently across cohorts of different scales.

Problem

• skin lesion segmentation
• skin lesion classification
• multi-task learning
• pattern analysis
• dermoscopy

Innovation

• Multi-task learning
• Pattern analysis priors
• Spatial alignment regularization
• Hybrid CNN-Transformer
• Skin lesion analysis