This repository contains the code that has been used in the study described here. We proposed for the first time an extensive comparison of blood vessel segmentation methods in Optical Coherence Tomography Angiography (OCTA). Each folder represents a vessel enhancement method (Frangi, Gabor, SCIRD-TS, OOF, CNN, and UNet), plus measuremnets to compute quality of teh segmentation (CAL, TopS). CS-Net implementation can be found here Folders contain README.md when necessary or comments to guide you through the code.
The code contains scripts in matlab and python (jupyter notebook).
Matlab: MATLAB R2018b
Python: 3.6.9 (tensorflow-gpu = 1.10.0, tensorflow = 1.12.0)
Implementations of Frangi, Gabor, SCIRD-TS, OOF have been already provided somewhere else. Info can be found in each folder.