Automated segmentation and classifcation of retinal features for glaucoma diagnosis

Glaucoma is a progressive optic neuropathy, which damages the optic nerve head and causes irreversible visual field loss. It is considered as one of the major cause of blindness. Clinical diagnosis of glaucoma involves fundus photography for examining the changes occurring due to glaucoma. Fundus images are capable of being processed by computational algorithms. Thus, development of an automated diagnostic system using image processing techniques is of great importance for detection of glaucoma during mass screening. The proposed method aims to develop a computer aided diagnostic (CAD) system for glaucoma detection. First, automated segmentation algorithms for optic disc and optic cup are developed which overcomes the reduced variability present between the region of interest and the background. Second, the segmented regions are used to obtain the clinical and textural features. Finally an efficient classification model is developed by considering the dynamic classifier selection methods. The proposed method is tested on a hospital dataset and publically available Drishti dataset. The quantitative results proves the efficiency of the adopted methodology and thus, can be incorporated in CAD of glaucoma.

Automated summary

The study aims to develop a computer aided diagnostic system for glaucoma detection, utilizing automated segmentation algorithms for optic disc and optic cup, extracting clinical and textural features for classification, and achieving effective results on hospital and public datasets.