Journals Proceedings

Abstract

Modification of the Adaptive Morphology and Mask Overlapping is proposed in this research for the blood vessel segmentation of the retinal image. The method consists of five main stages, first, enhancement of retinal image using adaptive enhancement method. Second stage is bottom Hat transformation for the segmentation and noise removal. Third, adaptive thresholding process to remove the remain noise in the image. Morphological operation is applied in the fourth stage, i.e. skeletonization and pruning. Skeletonization is used for the blood vessel thinning process, and pruning is used is to remove unused branches of the segmented blood vessel. The last stage in this research is noise removal using mask overlapping. Experiment is done using the retinal image from the DRIVE dataset. The sensitivity, specificity, and the accuracy of the segmentation method are 78.5%, 95.67%, and 87.12% respectively.