Retinal Vessels Segmentation by Improving Salient Region Combined with Sobel Operator Condition

Abstract

Medical images contribute greatly to help physicians identify abnormalities in the patient’s body in today’s health care. Retinal vessels are one of the effective methods for diagnosing diseases, such as: age-related macular degeneration, diabetes, hypertension, arteriosclerosis. However, manual analysis for retinal images is time-consuming and costly for ophthalmologists. In this paper, we proposed an approach for segmentation in retinal vessels by improving salient region map combined with Sobel mask. The algorithm includes two steps: superpixel detection and segmentation based on salient region map. The result of proposed method is better than the other methods.