Automated Retinal Vessel Analysis: A Novel Model for Blood Vessel Detection, Segmentation, and Clinical Characteristic Quantification

This research enlightens into a fully automated model for segmenting blood vessels and extracting clinical features from retinal fundus images, which is vital for diagnosing ophthalmologic and cardiovascular diseases. The model utilises contrast enhancement, noise reduction, edge detection, and a method for reconnecting blood vessel branches. It extracts properties like tortuosity, girth, and length of blood vessels. The model's performance surpasses existing approaches, achieving high accuracy, sensitivity, specificity, and positive and negative prediction values on DRIVE and HRF datasets. This innovation offers the potential to reduce specialists' workload, enhance diagnostic accuracy, and streamline the analysis of complex fundus images, ultimately improving patient care.