Diabetic Retinopathy Detection & Segmentation Using Color Image Processing

Abstract

Kaushick Parui, B. K. Tripathy

Diabetic Retinopathy has become a serious issue in the recent past being one of the primary causes for the loss of vision in diabetic patient. The primary motive of this project is to develop an automated technique for the detection of diabetic retinopathy using fundus images. This project involves the processing of the fundus images especially for the detection of diabetic retinopathy which is confirmed by the presence of microaneurysms, exudates, etc. Here, a novel approach has been adopted to perform image segmentation of all these visible stages in case of diabetic retinopathy and not only determine the stage of diabetic retinopathy but also give some more insights related to the disease. This project is completely done by utilizing the tools of Color Image Processing Viz. Segmentation technique, CLAHE technique, morphological operations, etc. Later, a comparative test was performed to test the correctness of the algorithm using various Diabetic Retinopathy Fundus Images & Normal Fundus Images. The algorithm seems to be at par with other approaches used in the modern world today to solve the problem of Diabetic Retinopathy and enables speedy diagnosis by doctors.

PDF

Share this article

Get the App