Influence of Krawtchouk and Charlier moment’s parameters on image reconstruction and classification

Abstract

The discrete orthogonal moments such as Krawtchouk and Charlier moments have been used successfully in the field of image analysis. These moments are based on discrete orthogonal polynomials that depend on the locality parameters (p for Krawtchouk and a1 for Charlier). However, the choice of the value of these parameters during the image analysis remains a problem without solutions. In the absence of an analytical study to make these choices, we will study experimentally, in this article, the influence of the variation of the parameters of discrete orthogonal polynomials on the analysis and in particular on the reconstruction and classification of the image. The results of the simulation show the influence of the variation of discrete orthogonal moments parameters of Krawtchouk and Charlier on the quality of reconstruction of the binary, grayscale and colors images in order to make the choice of the numerical value of each parameter. This allows us to obtain a range of parameter values that will be exploited in the classification task of image forms using the invariant moments of Krawtchouk and Charlier.