Machine Learning
Healthcare
Detection and localization of exudates in Diabetic Retinopathy fundus images
Project Principal Investigator(s): Neelam Sinha
Diabetic Retinopathy, a dis-order of the eye, is a consequence of Diabetes, that could lead to complete blindness. Hence early detection of the disease is very important. In a society like ours, the ratio of the population that suffers from Diabetes to the number of skilled specialists available for consultation is adversely skewed, making it very hard to provide the required care. In such a scenario, introducing systems that can automate screening of the disease would be very useful. In this project, we propose to build a system that detects whether or not the eye, whose fundus image is processed, is afflicted with Diabetic Retinopathy. We propose to build the system using the Deep Learning framework, that has demonstrated success over diverse problems in computer vision tasks such as classification, segmentation and object detection. Besides, we propose to utilize the idea of class-specific fixations to localize the pixels that represent exudates, the abnormality caused in the eye that suffers from Diabetic Retinopathy.
