Automated Optic Nerve Head Detection Based on Different Retinal Vasculature Segmentation Methods and Mathematical Morphology

• Main Authors: Tavakoli, Meysam, Nazar, Mahdieh, Golestaneh, Alireza, Kalantari, Faraz
• Format: Conference Proceeding
• Language: English
• Subjects: Biomedical imaging; Canny edge detector; Detectors; Diabetic retinopathy; Image edge detection; image processing; Image segmentation; Laplacianof-Gaussian edge detector; Match filter; Optic Nerve Head; Retina; retinal blood vessel; Transforms
• ISSN: 2577-0829
• DOI: 10.1109/NSSMIC.2017.8532764

Computer vision and image processing techniques provide important assistance to physicians and relieve their work load in different tasks. In particular, identifying objects of interest such as lesions and anatomical structures from the image is a challenging and iterative process that can be done by using computer vision and image processing approaches in a successful manner. Optic Nerve Head (ONH) detection is a crucial step in retinal image analysis algorithms. The goal of ONH detection is to find and detect other retinal landmarks and lesions and their corresponding diameters, to use as a length reference to measure objects in the retina. The objective of this study is to apply three retinal vessel segmentation methods, Laplacian-of-Gaussian edge detector, Canny edge detector, and Matched filter edge detector for detection of the ONH either in the normal fundus images or in the presence of retinal lesions (e.g. diabetic retinopathy). The steps for the segmentation are as following: 1) Smoothing: suppress as much noise as possible, without destroying the true edges, 2) Enhancement: apply a filter to enhance the quality of the edges in the image (sharpening), 3) Detection: determine which edge pixels should be discarded as noise and which should be retained by thresholding the edge strength and edge size, 4) Localization: determine the exact location of an edge by edge thinning or linking. To evaluate the accuracy of our proposed method, we compare the output of our proposed method with the ground truth data that collected by ophthalmologists on retinal images belonging to a test set of 120 images. As shown in the results section, by using the Laplacianof-Gaussian vessel segmentation, our automated algorithm finds 18 ONHs in true location for 20 color images in the CHASE-DB database and all images in the DRIVE database. For the Canny vessel segmentation, our automated algorithm finds 16 ONHs in true location for 20 images in the CHASE-DB database and 32 out of 40 images in the DRIVE database. And lastly, using matched filter in the vessel segmentation, our algorithm finds 19 ONHs in true location for 20 images in CHASE-DB database and all images in the DRIVE.