COLOR EDGE DETECTIONS BASED ON CELLULAR NEURAL NETWORK

Abstract

Color edge detection is one of the most important steps for RGB image recognition. In this paper, we first present two robustness design theorems for the Edgegray Detection Cellular Neural Network (EDGE CNN) and the counter detection (CD) CNN. Second, based on a color plane transform of RGB image and the two theorems, we design a color EDGE CNN and a CD CNN. As applications, the two CNNs detect successfully the edges of a standard color edge test pattern, and two popular RGB images, respectively. Our findings show that CNN may provide a useful tool for color image processing.