This paper presents a computational model to extract, from eye fundus images, the retinal vasculature and then to detect its features such as bifurcations and crossover points of retinal vessels. In particular, our approach can be articulated in six steps. The first five steps are represented by a combined application of five operators: Naka-Rushton filter, cluster filter, hyperbole filter, median filter and a skeleton process. These computational steps process retinal images in order to remove noise and then they can produce an optimised skeleton version of the vessels. Last step consists of detecting bifurcation and crossover points of the retinal vessels. These extracted features are a powerful tool for medical and biological evaluations of ocular and nonocular illness diagnoses.
|Titolo:||A Combined Method to Detect Retinal Fundus Features|
|Data di pubblicazione:||2005|
|Nome del convegno:||IEEE European Conference on Emergent Aspects in Clinical Data Analysis, EACDA 2005|
|Appare nelle tipologie:||4.1 Contributo in Atti di convegno|