A compressive sensing based efficient approach for early automatic detection of diabetic retinopathy using fundus image

Abstract

Diabetic Retinopathy (DR) is an abnormality related with eye. It is caused due to alteration of the blood vessels present in the retina. If this disease is not treated at the early time, it may lead to permanent vision loss. Numerous image processing techniques including Enhancement, Segmentation, Morphology, Image Fusion, Classification and Registration has been enveloped for the early recognition of DR through features such as blood vessels, exudes, haemorrhages and micro-aneurysms. This work deals with a new approach for the analysis of proliferate and non-proliferate DR with the help of Compressive Sensing (CS) technique. CS is a sparse way of effective reconstruction of signals that works mainly on two constraints namely sparsity and incoherence. In this work CS based various approaches are applied to fundus image. The results are studied and the best approach is used for Image Segmentation, Feature Extraction and Classification. newlineThe preliminary indicators of diabetic eye disease are small vessel enlargements known as micro-aneurysms. The automatic detection of DR by applying the fundus images of the retina to specific lesions of diabetic retinopathy is achieved in this research work. The pixel size of the fundus image is reduced using CS based approach in the pre-processing module thereby improving time complexity in Automatic detection of Diabetic Retinopathy with the help of CS. The CS proved the performance of the algorithm by reducing the image size without affecting the information. Identification of the blood vessels, micro-aneurysms and exudates has been achieved through learning-based approach newline