1.配置python环境

PyTorch使用2.1版本即可，其余的缺啥装啥(缺少的库可以等后续执行代码报错的时候再安装)。

2.下载数据集

下载BUSI等数据集，并按照如下示例存放：

data
└── BUSI
    ├── images
    └── masks

3. 数据集预处理

利用"./process/trans_BUSI.py"对BUSI数据集进行预处理，得到train.txt、val.txt和test.txt：

data
└── BUSI
    ├── images
    ├── masks
    ├── train.txt
    ├── val.txt
    └── test.txt

4. Train

在train.py中修改模型及相关参数后直接执行train.py文件即可。

Cite

@ARTICLE{11104828,
  author={Xiong, Youqiang and Shu, Xiu and Liu, Qiao and Yuan, Di},
  journal={IEEE Transactions on Consumer Electronics}, 
  title={HCMNet: A Hybrid CNN-Mamba Network for Breast Ultrasound Segmentation for Consumer Assisted Diagnosis}, 
  year={2025},
  volume={71},
  number={3},
  pages={8045-8054},
  keywords={Feature extraction;Image segmentation;Transformers;Computational modeling;Computer architecture;Adaptation models;Wavelet transforms;Noise;Decoding;Accuracy;Breast ultrasound images segmentation;hybrid network;wavelet feature;adaptive feature fusion},
  doi={10.1109/TCE.2025.3593784}}