Detecting the Contours via a New Approximation of the Gradient
Mohamed Lagzouli, Mustapha Rachidi, Youssfi Elkettani · International Journal of Computer Applications · 2014
Contours detection is a key component of many image processing and computer vision.This paper proposes and validates a new efficient gradient method for detecting the contours in grayscale image.This method is based on the average of two derivatives, obtained from two different steps.This mathematical formulation, derived from a discrete numerical differentiation of image, plays a central role in this method.There are presented some operators and mask of discrete functions, which are effective for the detection of contours.Comparison of the mask obtained using the two derivatives operator, with the usual linear masks, allows us to show the efficiency of the new mask.