mobarakol / Robustness_Stress_Testing

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Robustness_Stress_Testing

Instruction to run the code and reproduce the results of our paper "Robustness Stress Testing in Medical Image Classification".

Dataset

The CheXpert imaging dataset can be downloaded from https://stanfordmlgroup.github.io/competitions/chexpert/.
ISIC skin lesion dataset can be downloaded from https://github.com/GalAvineri/ISIC-Archive-Downloader.

Training command:

python train.py --img_data_dir '<path_to_data>/CheXpert-v1.0/'

To run our robustness stress testing

python perturbation.py --img_data_dir '<path_to_data>/CheXpert-v1.0/' --subgrp all

Citation

If you use this code for your research, please cite our paper.

@article{islam2023robustness,
  title={Robustness Stress Testing in Medical Image Classification},
  author={Islam, Mobarakol and Li, Zeju and Glocker, Ben},
  journal={MICCAI-MedAGI Workshop},
  year={2023}
}

About


Languages

Language:Jupyter Notebook 99.4%Language:Python 0.6%