FIGURE 2.
Quantification of macular microvasculature of a healthy and unhealthy retina using the deep-learning algorithm. Enface images of a healthy and unhealthy retina from the swept-source optical coherence tomography angiography (SS-OCTA) were converted into a skeletonized image. Fractal dimensions using Dbox and microvascular tortuosity were extracted from the skeletonized image using MATLAB.