zkyseu / cssnet

Official code for CSSNet: Cascaded Spatial Shift Network for Medical Image Segmentation

CSSNet

Official code for CSSNet: Cascaded Spatial Shift Network for Multi-organ Segmentation. Our paper has been accepted by Computers in Biology and Medicine!

Our project is based on TransUNet and AS-MLP. Thanks for their great work!

Installation

1. First install the pytorch.

conda create -n mlp python=3.8 -y
conda activate mlp
conda install pytorch==1.7.1 torchvision==0.8.2 torchaudio==0.7.2 cudatoolkit=10.2 -c pytorch

2. Following the TransUNet to install the dependency.
3. If you need the preprocessed data of Synapse dataset, please send an email to [email protected].

Training and Testing

1. Following the TransUNet to prepare the data.

2. Run the following code to train the model.

python train.py --dataset Synapse

3. You can use the following code to test the model.

python test.py --dataset Synapse

Core code

1. You can find the Cascaded-MLP code in as_mlp.py.

2. The network code is in Seg.py.

Citation

If you think our work is helpful, please cite our paper

@article{2024cssnet,
author = {Yeqin Shao, Kunyang Zhou, and Lichi Zhang},
title = {CSSNet: Cascaded spatial shift network for multi-organ segmentation},
journal = {Computers in Biology and Medicine},
year = {2024}
}