Hybrid clustering method for optic disc segmentation and feature extraction in retinal images

Optic Disc (OD) is one of the significant anatomical structure of human eye. The abnormalities change the size and shape of the human eye and hence the OD can play a significant role in the automated disease diagnostic system. The accuracy of the segmentation process and the feature extraction process must be exceedingly high to ensure the success of the system. In this work, emphasis is given to the segmentation of OD and feature extraction from OD. A hybrid method involving the fuzzy and level set method is used for segmentation. An extensive set of features is also extracted from these OD. Experimental results suggest promising results for the proposed approaches.