Statistical Analysis for Performance Evaluation of Image Segmentation Quality Using Edge Detection Algorithms

Edge detection is the most important feature of image processing for object detection, it is crucial to have a good understanding of edge detection algorithms/operators. Computer vision is rapidly expanding field that depends on the capability to perform faster segments and thus to classify and infer images. Segmentation is central to the successful extraction of image features and their ensuing classification. Powerful segmentation techniques are available; however each technique is ad hoc. In this paper, the computer vision investigates the sub regions of the composite image, brings out commonly used and most important edge detection algorithms/operators with a wide-ranging comparative along with the statistical approach. This paper implements popular algorithms such as Sobel, Roberts, Prewitt, Laplacian of Gaussian and canny. A standard metric is used for evaluating the performance degradation of edge detection algorithms as a function of Peak Signal to Noise Ratio (PSNR) along with the elapsed time for generating the segmented output image. A statistical approach to evaluate the variance among the PSNR and the time elapsed in output image is also incorporated. This paper provides a basis for objectively comparing the performance of different techniques and quantifies relative noise tolerance. Results shown allow selection of the most optimum method for application to image.