Hybrid Canny Zerocross Method for Edge Detection in Retina Identification Cases

Abstract

Edge detection is fundamental to Figure processing. Edges include much information in a figure, including the object's location, shape, size, and information about its texture. Since edge detection is a critical component of Figure processing for object detection, comprehend algorithms for edge detection. This is because the edges define an item's contours, serve as a demarcation between the object and its backdrop, and serve as a demarcation between overlapping objects. That is, if the edges of an image can be identified accurately, all things can be found. The proposal of this paper is the use of the Canny Zerocross hybrid method to perform better edge detection based on comparative studies and the incorporation of the Canny way, which is considered one of the best edge detection methods, with the Zerocross way (cross zero) which is a derivative of the laplacian. In this paper, the research data used is the retinal image dataset—data obtained from STARE (Structured Analysis of the Retina). The Veterans Administration Medical Center in San Diego and the Shiley Eye Center (ECS) at the University of California provided Figures and clinical data from the retinal images. The experimental results of the comparative study show that the Zerocross edge detection technique is better than the Canny edge detection technique. Meanwhile, edge detection and image identification would be better when combining the two methods (hybrid) based on merging studies.