PyJulie / SALL

official implementation for IEEE JBHI paper 'Synergic Adversarial Label Learning for Grading Retinal Diseases via Knowledge Distillation and Multi-task Learning'

SALL

Official implementation for IEEE JBHI paper 'Synergic Adversarial Label Learning for Grading Retinal Diseases via Knowledge Distillation and Multi-task Learning'

Please cite:

@article{ju2021synergic,
title={Synergic Adversarial Label Learning for Grading Retinal Diseases via Knowledge Distillation and Multi-task Learning},
author={Ju, Lie and Wang, Xin and Zhao, Xin and Lu, Huimin and Mahapatra, Dwarikanath and Bonnington, Paul and Ge, Zongyuan},
journal={IEEE Journal of Biomedical and Health Informatics},
year={2021},
publisher={IEEE}
}

This work uses a private datasets. You can find some useful dataset here.

Also, you can try a cifar-10 (5/5) dataset as a toy experiment. Our methods can also achieve improvments on those classes with similar features.

	Task A (1-5)	Task B (6-10)
Single Task	91.80	95.84
Ours	92.70	96.60

To Do

Pytorch implementation.