Preprocessed data for this experiment were provided by Charit´e University Hospital Berlin.
Data consist of the images of 53 healthy eyes. Each eye contains 47-49 scans (images) which makes 2597 images in total.
We are not allowed to upload data and use outside of the class.
We have compared 2 methods, which were proposed for medical image segmentation, namely UNet and CNN-S.
UNet code was forked from https://github.com/milesial/Pytorch-UNet
UNet:
Arslan Gait - Computer Science, Nazarbayev University, [email protected]
Aldiyar Bolatov - Department of Electrical and Computer Engineering, Nazarbayev University, [email protected]
Aslan Ubingazhibov - Computer Science, Nazarbayev University, [email protected]
Islambek Temirbek - Department of Electrical and Computer Engineering, Nazarbayev University,[email protected]
Shah, Abhay, et al. "Multiple surface segmentation using convolution neural nets: application to retinal layer segmentation in OCT images." Biomedical optics express 9.9 (2018): 4509-4526.
Ronneberger, Olaf, Philipp Fischer, and Thomas Brox. "U-net: Convolutional networks for biomedical image segmentation." International Conference on Medical image computing and computer-assisted intervention. Springer, Cham, 2015.