arXiv 2505.152342025-10-17eess.IVcs.AIcs.CV

用于医学图像分割的具有自适应Mamba类注意力与因果共振学习的UNet

UNet with Self-Adaptive Mamba-Like Attention and Causal-Resonance Learning for Medical Image Segmentation

Faculty of Computing and IT (FCIT)（计算与信息技术学院）；Sohar University（索哈尔大学）；Department of Intelligent Systems（智能系统系）；KTH Royal Institute of Technology（皇家理工学院）；College of Computer and Information Sciences（计算机与信息科学学院）；Imam Mohammad Ibn Saud Islamic University (IMSIU)（伊玛目穆罕默德·本·沙特伊斯兰大学）；University of Auckland（奥克兰大学）；Department of Computer Science, Faculty of Science（科学学院计算机科学系）；Northern Border University（北方边境大学）

Saqib Qamar, Mohd Fazil, Parvez Ahmad, Shakir Khan, Abu Taha Zamani

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

提出SAMA-UNet，通过自适应Mamba类注意力与因果共振多尺度模块，在医学图像分割中兼顾效率与精度，并在多个数据集上超越现有方法。

摘要：

医学图像分割在各种临床应用中扮演着重要角色；然而，现有的深度学习模型在效率和准确性之间面临权衡。卷积神经网络（CNNs）能很好地捕捉局部细节，但会遗漏全局上下文，而Transformer能处理全局上下文，但计算成本高昂。近来，状态空间序列模型（SSMs）显示出以线性复杂度捕捉长距离依赖的潜力；然而，由于与图像结构不兼容以及自回归假设，它们直接用于医学图像分割仍然受限。为克服这些挑战，我们提出了SAMA-UNet，一种新颖的U形架构，引入了两项关键创新。首先，自适应性类Mamba聚合注意力（SAMA）模块通过动态注意力加权自适应地整合局部和全局特征，从而高效地表示复杂的解剖模式。其次，因果共振多尺度模块（CR-MSM）通过调整特征分辨率和跨尺度的因果依赖来改善编码器-解码器交互，增强低层和高层特征之间的语义对齐。在MRI、CT和内窥镜数据集上的大量实验表明，SAMA-UNet始终优于基于CNN、Transformer和Mamba的方法。它在BTCV上达到85.38%的DSC和87.82%的NSD，在ACDC上达到92.16%和96.54%，在EndoVis17上达到67.14%和68.70%，在ATLAS23上达到84.06%和88.47%，跨模态建立了新的基准。这些结果证实了SAMA-UNet在结合效率与准确性方面的有效性，使其成为真实世界临床分割任务的一个有前景的解决方案。源代码可在GitHub上获取。

英文摘要：

Medical image segmentation plays an important role in various clinical applications; however, existing deep learning models face trade-offs between efficiency and accuracy. Convolutional Neural Networks (CNNs) capture local details well but miss the global context, whereas transformers handle the global context but at a high computational cost. Recently, State Space Sequence Models (SSMs) have shown potential for capturing long-range dependencies with linear complexity; however, their direct use in medical image segmentation remains limited due to incompatibility with image structures and autoregressive assumptions. To overcome these challenges, we propose SAMA-UNet, a novel U-shaped architecture that introduces two key innovations. First, the Self-Adaptive Mamba-like Aggregated Attention (SAMA) block adaptively integrates local and global features through dynamic attention weighting, enabling an efficient representation of complex anatomical patterns. Second, the causal resonance multi-scale module (CR-MSM) improves encoder-decoder interactions by adjusting feature resolution and causal dependencies across scales, enhancing the semantic alignment between low- and high-level features. Extensive experiments on MRI, CT, and endoscopy datasets demonstrate that SAMA-UNet consistently outperforms CNN, Transformer, and Mamba-based methods. It achieves 85.38% DSC and 87.82% NSD on BTCV, 92.16% and 96.54% on ACDC, 67.14% and 68.70% on EndoVis17, and 84.06% and 88.47% on ATLAS23, establishing new benchmarks across modalities. These results confirm the effectiveness of SAMA-UNet in combining efficiency and accuracy, making it a promising solution for real-world clinical segmentation tasks. The source code is available on GitHub.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。