By Pereira, C.; Veiga, D.; Mahdjoub, J.; Guessoum, Z.; Gon?alves, L.; Ferreira, M&pe
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Microaneurysms are the diabetic retinopathy rst sign and its early detection is crucial for blindness prevention. Several approaches can be found in literature for the automatic microaneurysm segmentation, but none of them has shown the required performance. In this study, a new approach is proposed based on an organization of agents enabling microaneurysms segmentation. This multiagent model is preceded by a preprocessing phase to allow the environment construction where agents are situated and interact. Then, microaneurysms segmentation emerges from agent interaction. With this study, competitive results comparing to more traditional algorithms were achieved, specially in detecting microaneurysms close to vessels. It seems that a very e cient system for the diabetic retinopathy diagnosis can be built using MAS mechanisms.