Improvement of Optic Disc Localization using Gabor Filters

The paper presents a supervised technique for the detection and localization of the optic disc (OD) in retinal images. The proposed processing technique is based on Discrete Fourier Transform (DFT) and Gabor filters (GFs). The algorithm of image patch processing and classification has two phases: the learning phase for the OD class definition and the testing phase for the patch processing and classification. Two features are used to check if a patch contains the OD: the magnitude and the phase values computed on the result of the convolution between the DFT of the patch and the bank of Gabor filters. Over 100 images from MESSIDOR database were tested and comparing with other similar works. The proposed algorithm gave better results in terms of accuracy of the OD localization for all types of OD.