CheSS: Chest X-ray pre-trained model via Self-Supervised contrastive learning

This is a PyTorch implementation of the CheSS paper:

@article{Cho2023,
doi = {10.1007/s10278-023-00782-4},
url = {https://doi.org/10.1007/s10278-023-00782-4},
year = {2023},
month = jan,
publisher = {Springer Science and Business Media {LLC}},
author = {Kyungjin Cho and Ki Duk Kim and Yujin Nam and Jiheon Jeong and Jeeyoung Kim and Changyong Choi and Soyoung Lee and Jun Soo Lee and Seoyeon Woo and Gil-Sun Hong and Joon Beom Seo and Namkug Kim},
title = {{CheSS}: Chest X-Ray Pre-trained Model via Self-supervised Contrastive Learning},
journal = {Journal of Digital Imaging}
}

Requirements

pip install -r requirements.txt

Pretrained model weight

Google Drive

model=resnet50(num_classes=1000)
pretrained_model="CheSS pretrained model path"ifpretrained_modelisnotNone:
ifos.path.isfile(pretrained_model):
print("=> loading checkpoint '{}'".format(pretrained_model))
checkpoint=torch.load(pretrained_model, map_location="cpu")
# rename moco pre-trained keysstate_dict=checkpoint['state_dict']
forkinlist(state_dict.keys()):
# retain only encoder_q up to before the embedding layerifk.startswith('module.encoder_q') andnotk.startswith('module.encoder_q.fc'):
# remove prefixstate_dict[k[len("module.encoder_q."):]] =state_dict[k]
# delete renamed or unused kdelstate_dict[k]
msg=model.load_state_dict(state_dict, strict=False)
assertset(msg.missing_keys) == {"fc.weight", "fc.bias"}
print("=> loaded pre-trained model '{}'".format(pretrained_model))
else:
print("=> no checkpoint found at '{}'".format(pretrained_model))
##freeze all layers but the last fcforname, paraminmodel.named_parameters():
ifnamenotin ['fc.weight', 'fc.bias']:
param.requires_grad=Falsemodel.fc=nn.Linear(2048, num_class)

or you can use gdown in Python

!gdownhttps://drive.google.com/uc?id=1IfiuQdKV7en9DFaB0NqNdsDkVbdyoVyD

Contact

Page: https://mi2rl.co

Email: kjcho@amc.seoul.kr