Retinal Blood Vessels Segmentation using Gabor Filters

Blood vessels are important parts of the anatomy of the human retina, any change in the structure of the vessels can be an indicator of retinal disorders. Early detection of such disorders can affect the treatment procedure, avoiding severe side effects like blindness. In this paper, a retinal eye blood vessels segmentation algorithm is proposed to aid the process of retinal screening. The proposed algorithm applied pre-processing techniques including noise removal followed by contrast stretching. Gabor filter is used to enhance the appearance of blood vessels using the directional field method. After that, segmentation using a global threshold applied followed by morphological post-processing techniques. The performance of the proposed technique was evaluated using a publicly available database (DRIVE) and proved to be a match for the state-of-art algorithms as the results were promising.