Optic cup segmentation is a very important step in glaucoma detection, where untreated glaucoma leads to irreversible optic nerve head (ONH) damage and loss of vision. Thus, it is the second cause of blindness around the world. In this paper, manual thresholding level technique in optic cup segmentation is applied, evaluated, and compared with other segmentation techniques. The optic cup segmented by thresholding level 240 from the optic disc partas a region of interest (ROI), after vessel removal by the morphological operation to improve segmentation process, filtering, contrast equalization and finally boundary smoothed by dilating and eroding operation and border clearing. The algorithm was evaluated on DRISHTI-GS database, which has anoptic cup segmentation ground truth. The proposed cup segmentation method evaluate by 3 parameters, which are Dice coefficient (DSC), Jaccard coefficient and structural similarity (SSIM) and achieved best Dice coefficient 73%, Jaccard coefficient 60%, and structural similarity 93%. Because the human visual system is good at perceiving structure, the SSIM quality metric agrees more closely with the subjective quality score, therefore, the obtained results so near to the ophthalmologist segmentation judgment. It was concluded that an optic cup (OC) segmentation algorithm via manual thresholding technique is easy, fast, and inexpensive computationally and obtained good segmentation results and would help in early detection of glaucoma by the ophthalmologist.
