Scholarlink

<p>Accurate and effective skin lesion segmentation plays a crucial role in clinical diagnosis. Despite the success of recent deep learning methods, challenges such as irregular shapes, blurred boundaries, and low contrast remain. To address these issues, we propose a multi-scale feature fusion and hybrid attention network (MFFHA-Net). The model incorporates nonlinear spiking-based encoder–decoder fusion modules to enhance multi-scale feature integration. A hybrid attention module is introduced at the bottleneck to capture both global and local contextual information. In addition, a feature mapping function is designed to refine attention weights and improve detail representation. The experimental results on four publicly available datasets (ISIC 2016, 2017, 2018 and <InlineEquation ID="IEq1"> <EquationSource Format="TEX">\(\textrm{PH}^2\)</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mtext>PH</mtext> <mn>2</mn> </msup> </math></EquationSource> </InlineEquation>) show that MFFHA-Net achieves competitive performance, with Dice scores of 0.9338, 0.8806, 0.9212, and 0.9503, respectively. Code in: https://github.com/xhucr/MFFHA-Net.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multi-scale feature fusion and hybrid attention network based on nonlinear spiking neural model for skin disease segmentation

  • 期刊论文

  • 发表日期：

    2026年5月15日

查看全文
  • Rui Cai,
  • Hong Peng,
  • Rikong Lugu

摘要

Accurate and effective skin lesion segmentation plays a crucial role in clinical diagnosis. Despite the success of recent deep learning methods, challenges such as irregular shapes, blurred boundaries, and low contrast remain. To address these issues, we propose a multi-scale feature fusion and hybrid attention network (MFFHA-Net). The model incorporates nonlinear spiking-based encoder–decoder fusion modules to enhance multi-scale feature integration. A hybrid attention module is introduced at the bottleneck to capture both global and local contextual information. In addition, a feature mapping function is designed to refine attention weights and improve detail representation. The experimental results on four publicly available datasets (ISIC 2016, 2017, 2018 and \(\textrm{PH}^2\) PH 2 ) show that MFFHA-Net achieves competitive performance, with Dice scores of 0.9338, 0.8806, 0.9212, and 0.9503, respectively. Code in: https://github.com/xhucr/MFFHA-Net.