Project information
- Category: Retinal Optic Disc Detection
- Data: DRIVE, STARE, DIARETDB0 and DIRETDB1 datasets
- Project date: August, 2018
- Project URL: https://doi.org/10.1016/j.bbe.2018.08.003
Automated fuzzy optic disc detection algorithm using branching of vessels and color properties in fundus images
This study presents a novel OD localization method that leverages the color and shape characteristics of the OD, along with the convergence point of primary blood vessels. The approach involves two main stages: initially, an algorithm identifies the approximate vessel convergence to roughly locate the OD, and subsequently, three new features are introduced, feeding into a fuzzy logic controller (FLC) with input membership functions tailored to these features. When tested on the DRIVE, STARE, DIARETDB0, and DIRETDB1 datasets, the method demonstrated enhanced performance, achieving success rates of 100%, 91.35%, 90%, and 100%, respectively, while accurately identifying OD centers and contours within a reasonable processing time.