H-fusion SEG: dual-branch hyper-attention fusion network with SAM integration for robust skin disease segmentation

• Published in: Scientific reports Vol. 15; no. 1; pp. 33668 - 24
• Main Authors: El‑Shafai, Walid, Ali, Anas M., Alzaben, Nada, El‑Fattah, Ibrahim Abd
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 29.09.2025; Nature Publishing Group; Nature Portfolio
• Subjects: 631/114; 631/67; 639/166; 639/705; 692/699; Accuracy; Algorithms; Architecture; Artificial intelligence; Benchmarks; Boundaries; Datasets; Deep learning; Dermoscopy - methods; Dual-encoder architecture; Humanities and Social Sciences; Humans; Hyper-attention fusion; Image Interpretation, Computer-Assisted - methods; Image processing; Image Processing, Computer-Assisted - methods; Medical research; Morphology; multidisciplinary; Network topologies; Neural networks; Science; Science (multidisciplinary); Segment anything model (SAM); Semantics; Skin diseases; Skin Diseases - diagnosis; Skin Diseases - diagnostic imaging; Skin lesion segmentation; U-Net; Vision transformers; Skin lesion segmentation; Dual-encoder architecture; Vision transformers; Hyper-attention fusion; U-Net; Segment anything model (SAM)
• ISSN: 2045-2322; 2045-2322
• DOI: 10.1038/s41598-025-18202-8

Accurate dermoscopic lesion segmentation is challenging because existing methods struggle to preserve fine-grained local structures while capturing long-range semantic context, leading to reduced robustness against unclear boundaries, imaging artifacts, and dataset shifts. We propose Hyper-Fusion Segmentation (H-Fusion SEG), a dual-branch framework that combines a boundary-sensitive U-Net encoder–decoder with a Segment Anything Model branch to jointly extract high-resolution local details and robust global semantics. A novel hyper-attention fusion module adaptively integrates these heterogeneous features and is optimized with boundary-aware objectives to enhance delineation and interpretability. On the ISIC-2016 dataset, H-Fusion SEG achieves IoU = 0.8775 and Dice = 0.9269 (+ 1.28% IoU, + 1.38% Dice over baselines), and on ISIC-2018, it achieves IoU = 0.9329 and Dice = 0.9629 (+ 8.69% IoU, + 6.69% Dice over baselines), with strong generalization to the HAM10000 dataset. These gains are particularly pronounced for complex lesions with indistinct or ambiguous boundaries. The proposed framework offers a flexible and generalizable solution for medical image segmentation, with promising potential for precise and reliable computer-aided diagnostic tools in dermatology. Code is available at: https://github.com/AnasHXH/Skin-DiseaseS-Segmentation .