Blood Vessel Detection in Fundus Images by a Multi-agent Approach

Abstract

The segmentation of retinal blood vessels by digital color fundus images analysis is crucial for several medical diagnostic systems, such as the diabetic retinopathy early diagnosis. This pathology has been shown to be the most common cause of blindness among working age people in developed countries. Several interesting approaches have been done in segmenting the blood vessels by image processing techniques applied in fundus images, but none of them has shown the required performance to be applied in clinical practices. Therefore, a new approach is proposed based on an agents’ organization enabling vessels detection. This multi-agent approach is preceded by a preprocessing phase in which the fundamental filter is a Kirsch derivative improved version. This first phase allows an environment construction where agents are situated and interact. Then, blood vessel edges detection emerged from agent interaction. According to this study, competitive results as compared with those present in the literature were achieved. It seems to be that a very efficient system for diabetic retinopathy diagnosis can be built using MAS mechanisms.