A color edge detector based on Dempster-Shafer theory

Segmentation based on contour detection is a relevant stage before image interpretation or pattern recognition. This paper is concerned with color image filtering and color edge detecting. These 2 techniques utilize the Dempster-Shafer theory. After the description of color image filtering which generalizes Nagao's filter, a method for computing the orientation and magnitude of a gradient on color images is presented. Both filtering and edge detecting use a 5x5 window. Some choices in the algorithms permit to reduce the computing complexity of evidential theory. Finally, a method which extracts the local maxima of gradient is described.