MICROANEURYSM DETECTION USING ACTIVE CONTOUR MODEL WITH DIABETIC RETINOPATHY SCREENING

Digital Electronics

NATIONAL CONFERENCE ON TECHNOLOGICAL ADVANCEMENTS IN ENGINEERING

Diabetic retinopathy is most leading disabling diseases effected in retina, it will be the leading causes of preventable blindness in the world. For automated screening programs of diabetic retinopathy, image processing and analysis algorithms have to be developed. Microaneurysms are the first sign of diabetic retinopathy for automatic screening. The number of microaneurysms are used to indicate the severity of the disease like diabetic retinopathy mild, diabetic retinopathy moderate etc. Early microaneurysm detection can help to reduce the occurrence of blindness. Candidate objects are first located by applying a dark object filtering process, active contour model for detect the vasculature and neural network for classification. Contour model segmentation in fundus images is a nontrivial task due to variable size of vessels, relatively low contrast, and potential presence of pathologies like microaneurysms and other diabetic retinopathy signs. This model can also find out different stages of diabetic retinopathy like diabetic retinopathy mild, diabetic retinopathy moderate, diabetic retinopathy severe etc with high accuracy using neural network classifier. Diabetic retinopathy pathologies were further categorized into several groups. In this paper several different databases are presented. The proposed method is performed on the green channel of retinal fundus image. the MAs, hemorrhages and vessels normally present the highest contrast against the surrounding background in green channel, so microaneurysm appear as bright structure and get finer details [6]. MA detection process is divided into mainly 4 steps, there are preprocessing, candidate extraction through multilayered dark object filtering, vasculature detection based on active contour model and classification.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.