This repo covers the winning solutions from Team fdvts_mm in the MICCAI 2023 Myopic Maculopathy Analysis Challenge.
We download the datasets from MMAC2023 and DDR.
cd MMAC_task1
train a ResNet50 model
python main.py --challenge 1 --model resnet50 --visname resnet50
cd MMAC_task2
train segmentation model
bash train.sh
Ensemble appoaches in MMAC_task2/ensemble_model.py
If you use this code, please cite the following papers:
@inproceedings{hou2023towards,
title={Towards Label-Efficient Deep Learning for Myopic Maculopathy Classification},
author={Hou, Junlin and Xu, Jilan and Xiao, Fan and Zhang, Bo and Xu, Yiqian and Zhang, Yuejie and Zou, Haidong and Feng, Rui},
booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
pages={31--45},
year={2023},
organization={Springer}
}
@inproceedings{xiao2023ensemble,
title={Ensemble Deep Learning Approaches for Myopic Maculopathy Plus Lesions Segmentation},
author={Xiao, Fan and Hou, Junlin and Xu, Jilan and Xu, Yiqian and Zhang, Bo and Zhang, Yuejie and Zou, Haidong and Feng, Rui},
booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
pages={46--55},
year={2023},
organization={Springer}
}