Automated detection of optic disc using vessels tracking

Optic disc is an important anatomical feature of the retinal image and helps in the diagnosis of eye diseases. The proposed method is based on the characteristics of retinal image. An important characteristic of the retinal image is that blood vessels egress from the optic disc and become thin as they move away from the optic disc. If blood vessels are tracked and features are calculated then they end up on the optic disc. Given a retinal image the proposed method first enhances the blood vessels using 2-D Gabor wavelets and segmenting the blood vessels using iterative thresholding technique to join the vessels properly. The resultant image is thinned using morphological operation and matched with the 2-D Gaussian function by tracking the blood vessels. Finally Optic disc is localized in a window. The proposed method is evaluated using publicly available databases and results show that the proposed method identifies optic disc with high accuracy.