Official PyTorch Code for Hybrid Window Attention Based Transformer Architecture for Brain Tumor Segmentation: Solution for FeTS 2022 Task 2.
Download data from FeTS 2022 challenge page and put it under data folder.
Please prepare an environment with python=3.8, and then run the command "pip install -r requirements.txt" for the dependencies.
- Swin-T: https://github.com/SwinTransformer/storage/releases/download/v1.0.8/swin_tiny_patch4_window7_224_22k.pth
- Download Swin-T pre-trained weights and add it under pretrained_ckpt folder
- CR-Swin2-VT: https://drive.google.com/file/d/18ukXmZ5TzNgSUKco8PzPzPHUrkc510Z1/view?usp=sharing
- Download CR-Swin2-VT pre-trained model and add it under saved_model folder before running test.py
- Train : Run the train script on BraTS 2021/FeTS 2022 Training Dataset with Base model Configurations.
python train.py
- Test : Run the test script on FeTS 2022 Validation Dataset.
python test.py
This repository makes liberal use of code from open_brats2020.
@article{peiris2022hybrid,
title={Hybrid Window Attention Based Transformer Architecture for Brain Tumor Segmentation},
author={Peiris, Himashi and Hayat, Munawar and Chen, Zhaolin and Egan, Gary and Harandi, Mehrtash},
journal={arXiv preprint arXiv:2209.07704},
year={2022}
}