Official implementation of paper "Retinal Image Restoration and Vessel Segmentation using Modified Cycle-CBAM and CBAM-UNet"
Cycle-consistent Generative Adversarial Network (CycleGAN) with Convolutional Block Attention Module (CBAM) - Cycle-CBAM. Modified UNet with CBAM - CBAM-UNet.
-
Create anaconda environment with python=3.9.
-
Download DRIVE, STARE and CHASE DB1 in UNet folder.
-
Run aug_drive.py, aug_stare.py, aug_chase.py in UNet folder.
-
Run train.py in UNet folder selecting the dataset.
-
Run test.py in UNet folder selecting the dataset.
-
Download EyeQ dataset and place in datasets folder according to the qualitites: 0, 1, and 2.
-
Run train.py in main folder.
-
Run test.py in main folder.
If you use this code, please use the following BibTeX entry.
@inproceedings{alimanov2022retinal,
title={Retinal Image Restoration and Vessel Segmentation using Modified Cycle-CBAM and CBAM-UNet},
author={Alimanov, Alnur and Islam, Md Baharul},
booktitle={2022 Innovations in Intelligent Systems and Applications Conference (ASYU)},
pages={1--6},
year={2022},
organization={IEEE}
}