Hybrid Ensemble of Segmentation-Assisted Classification and GBDT for Skin Cancer Detection with Engineered Metadata and Synthetic Lesions from ISIC 2024 Non-Dermoscopic 3D-TBP Images

2025-06-03

AI Summary

Accurate benign-malignant classification of skin lesions from 3D full-body images captured under non-dermoscopic conditions (e.g., smartphone photography) remains challenging due to low contrast, occlusion, and severe class imbalance. Method: We propose a segmentation-guided gradient-boosting hybrid framework: (1) U-Net guides lesion localization; (2) multi-scale visual features are extracted by fusing EVA02 ViT and our novel EdgeNeXtSAC architecture; (3) a diagnosis-driven relabeling strategy standardizes heterogeneous multi-source annotations into three clinically consistent classes, while Stable Diffusion synthesizes high-fidelity malignant lesion samples to mitigate class bias; (4) patient-level relational graph features are engineered and fed into XGBoost/GBDT for interpretable clinical decision-making. Results: On the ISIC 2024 SLICE-3D benchmark, our method achieves a pAUC@TPR>80% of 0.1755—the highest score in this track—demonstrating substantial improvement in remote triage reliability under resource-constrained settings.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Skin cancer is among the most prevalent and life-threatening diseases worldwide, with early detection being critical to patient outcomes. This work presents a hybrid machine and deep learning-based approach for classifying malignant and benign skin lesions using the SLICE-3D dataset from ISIC 2024, which comprises 401,059 cropped lesion images extracted from 3D Total Body Photography (TBP), emulating non-dermoscopic, smartphone-like conditions. Our method combines vision transformers (EVA02) and our designed convolutional ViT hybrid (EdgeNeXtSAC) to extract robust features, employing a segmentation-assisted classification pipeline to enhance lesion localization. Predictions from these models are fused with a gradient-boosted decision tree (GBDT) ensemble enriched by engineered features and patient-specific relational metrics. To address class imbalance and improve generalization, we augment malignant cases with Stable Diffusion-generated synthetic lesions and apply a diagnosis-informed relabeling strategy to harmonize external datasets into a 3-class format. Using partial AUC (pAUC) above 80 percent true positive rate (TPR) as the evaluation metric, our approach achieves a pAUC of 0.1755 -- the highest among all configurations. These results underscore the potential of hybrid, interpretable AI systems for skin cancer triage in telemedicine and resource-constrained settings.

Problem

• Classify malignant and benign skin lesions using hybrid AI
• Address class imbalance with synthetic lesions and relabeling
• Improve skin cancer detection in non-dermoscopic smartphone-like conditions

Innovation

• Hybrid ensemble of segmentation-assisted classification and GBDT
• Synthetic lesions generated with Stable Diffusion
• Vision transformers and convolutional ViT hybrid