HCMA-UNet: A Hybrid CNN-Mamba UNet with Axial Self-Attention for Efficient Breast Cancer Segmentation

TL;DR

This paper introduces HCMA-UNet, a lightweight hybrid CNN-Mamba network with axial self-attention for efficient and accurate breast cancer lesion segmentation in DCE-MRI, outperforming existing methods.

Contribution

The study presents a novel hybrid network architecture with a Multi-view Axial Self-Attention MISM module and a new loss function, achieving state-of-the-art results with fewer parameters.

Findings

Achieves superior segmentation accuracy on multiple datasets.

Maintains high efficiency with only 2.87M parameters.

Demonstrates good cross-architecture generalization.

Abstract

Breast cancer lesion segmentation in DCE-MRI remains challenging due to heterogeneous tumor morphology and indistinct boundaries. To address these challenges, this study proposes a novel hybrid segmentation network, HCMA-UNet, for lesion segmentation of breast cancer. Our network consists of a lightweight CNN backbone and a Multi-view Axial Self-Attention Mamba (MISM) module. The MISM module integrates Visual State Space Block (VSSB) and Axial Self-Attention (ASA) mechanism, effectively reducing parameters through Asymmetric Split Channel (ASC) strategy to achieve efficient tri-directional feature extraction. Our lightweight model achieves superior performance with 2.87M parameters and 126.44 GFLOPs. A Feature-guided Region-aware loss function (FRLoss) is proposed to enhance segmentation accuracy. Extensive experiments on one private and two public DCE-MRI breast cancer datasets demonstrate that our approach achieves state-of-the-art performance while maintaining computational efficiency. FRLoss also exhibits good cross-architecture generalization capabilities. The source code is available at https://github.com/Haoxuanli-Thu/HCMA-UNet.