THE FIRST ABSOLUTE CENTRAL MOMENT IN IMAGE ANALYSIS

摘要

In this paper we show how the generalization of the first absolute central moment gives rise to a class of nonlinear filters and how they can be used in image analysis to enhance lines, edges, corners and intersections between different discontinuities. Since the filters are nonlinear the recovered edge information can be also combined to obtain information that would not be obtained by varying the parameters of the original filter. Furthermore, we show how a mass center of the first absolute central moment can be defined and how this can be used to develop a new contour tracking procedure. The mass centers computed at the points of a given approximate starting contour are closer to the "true" contour than the points of the starting contour. Therefore, the final contour can be localized by iteratively computing the mass centers of the first absolute central moment.