A Robust Method for Finding Macular EDEMA Using GLCM Feature Extractor

A Robust Method for Finding Macular EDEMA Using GLCM Feature Extractor

Diabetic Macular Edema (DME) is the most common cause of blindness. We can avoid the visual impairment by detecting DME in its early stage. To assess the effects of sight threatening disease on human vision, a two-stage methodology is proposed. That is for the detection and classification of DME severity from color fundus images, before significant visual loss. DME detection is carried out via a supervised learning approach using the normal fundus images. Global characteristics of the fundus images are captured through GLCM feature extraction technique for discriminate normal image from the diseased one. An algorithm based on rotational symmetry of macular region examines the severity of disease. The proposed method is an effective and clinically viable technique for detecting diabetic DME before visual loss.