In this paper a contribution in the detection of diabetic symptoms using contrast enhancement of fundus images is proposed by synthesizing a hybrid neurofuzzy system via a sparsely-connected neural network. A fuzzy technique is firstly considered and a suitable coding of fuzzy rules is developed. A sparsely-connected neural network is then synthesized to provide enhanced contrast retinal images with bimodal histograms. Obtained contrast-enhanced images present pale areas that can more easily reveal suspect diabetic symptoms and enable accurate segmentations and measurements of the seriousness of diabetic symptoms. Experimental examples are reported.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.