README.md

SaB-Net: Self-attention backward network for gastric tumor segmentation in CT images

Overview

We provide the PyTorch implementation of our CIBM submission "SaB-Net".

Files description

SaB-Net contains the following folders:

├── data      \ Reference to data>README.md file for detail
          ├── SaB_processed     \ A folder to save the processed dataset
          ├── SaB_raw
                    ├── Dataset001_GTS     \ A raw data folder for a dataset
                    └── ......
          └── SaB_results     \ A folder to store the trained dataset results
└── src
          ├── execution     \ store the running scripts
          ├── network     \ store the model
          └── scripts     \ functional codes

Usage

Install requirements

pip install -r requirements.txt

Preprocess

python execution/preprocess.py -r [SaB_raw folder] -p [SaB_processed folder] -D [dataset_ID] 

Predict

python execution/predict.py -i [input_folder] -o [output_folder] -r [SaB_results folder] -d [cpu|gpu] -D [dataset_ID]

Train

Working on