IEEE 2016 International Conference on Wireless Communications, Signal Processing and Networking (WiSPNET 2016)

186 / 2015-12-03 19:34:19

Optic disc and optic cup segmentation for Glaucoma screening using Superpixel classification

8622,8623,8624,8625

For early detection of eye diseases automatic retinal image analysis is emerging as an important screening tool. Glaucoma is one of the most common causes of blindness in which neuro degeneration of the optic nerve takes place. One of standard procedure for detecting glaucoma is the manual examination of optic disk (OD). But this manual assessment is subjective, time consuming and costly procedure. Therefore, automatic assessment of optic nerve head seems to be very beneficial. We propose a glaucoma screening method by segmenting optic disc and optic cup using Superpixel classification. In this proposed approach, pre-processing such as image filtration, color contrast enhancement are performed on obtained fundus image which is followed by a combined approach for image segmentation and classification using texture, thresholding and morphological operation. To obtain accurate boundary delineation multimodalities including K-Means clustering, Gabor wavelet transformations are used. For cup segmentation, we incorporate previous knowledge of the cup by including location information. For glaucoma screening now CDR is computed based on the segmented disc and cup. System is tested on 27 retinal images from database, out of that 11 images gives result as moderate glaucoma, 12 gives normal result and 3 gives result as severe glaucoma, while 2 images shows incorrect reading. System achieves 98% of accuracy.