Feature Extraction for Early Detection of Diabetic Retinopathy

Automatic detection of lesions in retinal images can assist in early diagnosis and screening of Diabetic Retinopathy. Exudates are the primary sign of Diabetic Retinopathy. So detection of exudates is the fundamental requirement to diagnose the progress of Diabetic Retinopathy. The detection of optic disk is very essential to locate various features in retinal image. A robust and computational efficient approach for localization of different features and lesions in retinal images is presented in this paper. We propose a new constraint for optic disk detection, here we first detect the blood vessel and exudates patches and then we subtract both to get optic disk. Many of the features such as blood vessel, exudates and optic disk can be detected quite accurately using morphological operations applied appropriately.