CZS-Ma: Comprehensive Zigzag Scan Mamba Network for Precise and Efficient Segmentation

0. Main Environments

conda create -n your_environment_name python=3.8
conda activate your_environment_name
pip install torch==1.13.0 torchvision==0.14.0 torchaudio==0.13.0 --extra-index-url https://download.pytorch.org/whl/cu117
pip install packaging
pip install timm==0.4.12
pip install pytest chardet yacs termcolor
pip install submitit tensorboardX
pip install triton==2.0.0
pip install causal_conv1d==1.0.0  # causal_conv1d-1.0.0+cu118torch1.13cxx11abiFALSE-cp38-cp38-linux_x86_64.whl
pip install mamba_ssm==1.0.1  # mmamba_ssm-1.0.1+cu118torch1.13cxx11abiFALSE-cp38-cp38-linux_x86_64.whl
pip install scikit-learn matplotlib thop h5py SimpleITK scikit-image medpy yacs

Alternatively, you can use Autodl to quickly get started. Select state-spaces/mamba/mamba-ssm1.1.1_causal-convld1.1.3/v1 in the community mirror

1. Prepare the dataset

ISIC datasets

• The ISIC17 and ISIC18 datasets, divided into a 7:3 ratio, can be found here {Baidu or GoogleDrive}.

• After downloading the datasets, you are supposed to put them into './data/isic17/' and './data/isic18/', and the file format reference is as follows. (take the ISIC17 dataset as an example.)

• './data/isic17/'

  • train
    • images
      • .png
    • masks
      • .png
  • val
    • images
      • .png
    • masks
      • .png

Synapse datasets

• For the Synapse dataset, you could follow Swin-UNet to download the dataset, or you could download them from {Baidu}.

• After downloading the datasets, you are supposed to put them into './data/Synapse/', and the file format reference is as follows.

• './data/Synapse/'

  • lists
    • list_Synapse
      • all.lst
      • test_vol.txt
      • train.txt
  • test_vol_h5
    • casexxxx.npy.h5
  • train_npz
    • casexxxx_slicexxx.npz

2. Prepare the pre_trained weights

• The weights of the pre-trained RDZMamba could be downloaded here or Baidu. After that, the pre-trained weights should be stored in './pretrained_weights/'.

3. Train the CZS-Mamba

cd RDZMamba-Rapid-Diagonal-Zigzag-Mamba-Network-for-Medical-Image-Segmentation-master
python train.py  # Train and test RDZMamba on the ISIC17 or ISIC18 dataset.
python train_synapse.py  # Train and test RDZMamba on the Synapse dataset.

4. Obtain the outputs

• After trianing, you could obtain the results in './results/'

data download link:

https://drive.google.com/drive/folders/1KbVXgv4XnQ2dfGeq2NSWOBNDe-F32yrU?usp=sharing