A Comparative Study On Color Edge Detection

Several approaches of different complexity already exist to edge detection in color images. Nevertheless, the question remains of how different are the results when employing computational costly techniques instead of simple ones. This paper presents a comparative study on different approaches to color edge detection. The approaches are based on the Sobel operator, the Laplace operator, the Mexican Hat operator, different realizations of the Cumani operator, and the Alshatti-Lambert operator. Furthermore, we present an efficient algorithm for implementing the Cumani operator. All operators have been applied to several synthetic and real images. The results are presented in this paper. We show that the quality of the results increases if Gaussian masks of larger width are used in the derivation process instead of simple 3 x 3 masks as suggested in the underlying papers. Moreover, multiresolution approaches can be applied to color images when using Gaussian masks with different standard deviations in the edge detection scheme.