In clinical practice, medical image segmentation provides useful information
on the contours and dimensions of target organs or tissues, facilitating
improved diagnosis, analysis, and treatment. In the past few years,
convolutional neural networks (CNNs) and Transformers have dominated this area,
but they still suffer from either limited receptive fields or costly long-range
modeling. Mamba, a State Space Sequence Model (SSM), recently emerged as a
promising paradigm for long-range dependency modeling with linear complexity.
In this paper, we introduce a Large Window-based Mamba U}-shape Network, or
LMa-UNet, for 2D and 3D medical image segmentation. A distinguishing feature of
our LMa-UNet is its utilization of large windows, excelling in locally spatial
modeling compared to small kernel-based CNNs and small window-based
Transformers, while maintaining superior efficiency in global modeling compared
to self-attention with quadratic complexity. Additionally, we design a novel
hierarchical and bidirectional Mamba block to further enhance the global and
neighborhood spatial modeling capability of Mamba. Comprehensive experiments
demonstrate the effectiveness and efficiency of our method and the feasibility
of using large window size to achieve large receptive fields. Codes are
available at this https URL

通过利用大窗口进行局部的空间建模，同时保持超强的全局建模效果，在医学图像分割领域提出了基于大窗口的 Mamba U 形网络 (LMa-UNet)，通过设计新颖的分层双向 Mamba 块进一步增强了全局和邻域空间建模能力，全面实验验证了方法的有效性和高效性，并证明了使用大窗口大小实现大感受野的可行性。