In this paper a neurofuzzy network able to enhance contrast of retinal images for the detection of suspect diabetic symptoms is synthesized. Required fuzzy parameters are determined by ad hoc neural networks. Contrast-enhanced images are then segmented to isolate suspect areas by an adequate thresholding, which minimizes classification errors. In output images suspect diabetic regions are isolated. Capabilities and performances of the suggested network are reported and compared to scientific results.
A Neurofuzzy Network for Supporting Detection of Diabetic Symptoms / Carnimeo, Leonarda (LECTURE NOTES IN ELECTRICAL ENGINEERING). - In: Proceedings of the European Computing Conference. Volume 1 / [a cura di] Nikos Mastorakis; Valeri Mladenov; Vassiliki T. Kontargyri. - STAMPA. - Boston, MA : Springer, 2009. - ISBN 978-0-387-84813-6. - pp. 45-52 [10.1007/978-0-387-84814-3_6]
A Neurofuzzy Network for Supporting Detection of Diabetic Symptoms
Carnimeo, Leonarda
2009-01-01
Abstract
In this paper a neurofuzzy network able to enhance contrast of retinal images for the detection of suspect diabetic symptoms is synthesized. Required fuzzy parameters are determined by ad hoc neural networks. Contrast-enhanced images are then segmented to isolate suspect areas by an adequate thresholding, which minimizes classification errors. In output images suspect diabetic regions are isolated. Capabilities and performances of the suggested network are reported and compared to scientific results.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
