Blood Vessels Extraction of Retinal Image Using Morphological Operations

Blood vessel extraction is on high demand now-a-days as it plays very important role is diagnosing eye related disease in faster rate and efficient way as compared to manual segmentation and detection. The proposed algorithm focuses mainly on pre-processing, extraction and removing frills. Pre-processing has the main important role to play as the result is totally dependent on this step, so pre-processing is done by the use of Contrast Limited Adaptive Histogram Equalization (CLAHE). In the next step that is extraction of blood vessels is done through mathematical morphological functions as the vessels resemble patterns. The HRF database consisting of 45 images is used. The proposed algorithm is implemented in MATLAB 2013a software. The proposed method provides the accuracy of 95.95%, 96.26% and 95.37% for healthy, glaucoma and diabetic retinopathy cases respectively.