If you think this repo helps your research, please consider ref this paper (ACCV Workshop 2018, oral.) Thanks! A U-Net Segmentation is trained on the classical Drive (Utrecht University) dataset. (our model was released in 2017)
Georgia Tech, KAUST, U Waterloo Kyoto U
@inproceedings{yang2018auto,
title={Auto-classification of retinal diseases in the limit of sparse data using a two-streams machine learning model},
author={Yang, C-H Huck and Liu, Fangyu and Huang, Jia-Hong and Tian, Meng and Lin, MD I-Hung and Liu, Yi Chieh and Morikawa, Hiromasa and Yang, Hao-Hsiang and Tegn{\`e}r, Jesper},
booktitle={Asian Conference on Computer Vision},
pages={323--338},
year={2018},
organization={Springer}
}
Supplymentary 2019
python run_training.py
(a) Test Image (b) Ground Truth (c) Automatic Segementation after U-Net Image Model