On the analysis of contour figures by using cross correlation function

Abstract

It is necessary to evaluate the similarity of figures and to extract the features of figure in image processing. A contour figure can be processed by analyzing the complex valued function which presents the objective contour. The feature function is defined by cross correlation function of an objective contour figure and a standard contour figure. The feature function is also a complex valued function with the same period. The feature function indicates how a part of the contour is similar to a certain shape. We describe the similarity of contour figures by using the feature function. Various feature functions are produced by changing the standard contour. This paper reports the theory and experiment of this method.