A Review of the Classification of Artery and Vein Retinal Vessels Based on Machine learning.

The classification of artery and vein vessels in color retinal images is the basis of the detection, classification of the extent, and therapeutic evaluation of fundus lesions. Most changes in retinal artery and vein vessels in terms of caliber, width, and curvature are caused or affected by fundus pathology, and therefore, the classification of retinal artery and vein vessels can provide a reliable reference for later lesion diagnosis. In this paper, based on an in- depth study of the literature on retinal artery and vein vessel classification, we compare the currently used methods and summarize the advantages and disadvantages of these methods as well as the improvement methods.