Optical coherence tomography (OCT) blood vessel segmentation with CNN (U-Net).
Optical coherence tomography (OCT) image segmentation with a U-Net.
Followed this tutorial:
OCT scans were acquired manually through scanning subjects as part of other studies conducted by the Demer Ocular Motility Lab. Subjects included both controls and those with optic neuropathies.
Manual segmentation of the scans was done using Adobe Photoshop using the pencil tool.
Images were converted to grayscale (original scans) and binary (masks). Some of the manual segmentation were not completely binary, which was accounted for by setting the values to 0 and 1 in the code.
