Development of Edge Detection for Image Segmentation

Abstract

Abstract The edge detection technique is a fundamental phase of image segmentation. The purpose of the image segmentation algorithm is to distinguish the boundary of objects in different regions and it relies on discontinuities in image values between distinct regions. The objectives of this research are to a) develop an interface for image edge detection based on derivatives using MATLAB and b) measure the PSNR, SNR and MSE values for analysis based on experiments conducted. Results show that, Lena image produces PSNR values of 20.9 dB (Canny), 20.0 dB (Log), 20.1 dB (Prewitt), 20.0 dB (Sobel) and 20.0 dB (Robert). Meanwhile, MSE gives 80.5 dB (Canny), 83.1 dB (Log), 80.9 dB (Prewitt), 81.0 dB (Sobel) and 81.0 dB (Robert) after the edge detection process. The finding shows that Canny has given a winning performance in PSNR value and low in noise rate for JPEG type of image in image segmentation. Finally, the impact of edge detection techniques produces a better solution for image segmentation.