An optimized SVM classifier for detection of glaucoma by means of improved segmentation.

Glaucoma is a retinal disease with neuro degenerative persistent disorder that gradually damages the optic nerve and the world's leading source of blindness. Many applications in health sector massively utilize computer-vision based and Content-based image analysis practices to detect these chronic diseases. The fundus digital camera is focused to inspect structures like optic disc, retina, lens etc., from captured fundus images and aimed at detecting abnormalities in a human eye. The pathological disorder of the eye optical nerve due to Intra Ocular Pressure (IOP) will result to Glaucoma. Premature symptoms of glaucoma are non-observable and may result to partial vision loss and if left untreated, undiagnosed may lead to perpetual vision loss in future. Lot of human efforts are needed to spot this disease manually, which is not suggestible. Rapidly, many more studies have been evolved to automatically detect, diagnose and analyze the neurodegeneration infection called glaucoma using image processing, deep learning and computer vision processing techniques. These studies demonstrated that early discovery of glaucoma is best possible and can prevent patients from retinal impairment which decreases visual acuity. We report our finding with an enhanced framework with improved segmentation, modified SVM classification to classify retinal fundus images, and attaining accuracy of 97% when compared to state-of-the-art methods. [ABSTRACT FROM AUTHOR]