Classification of point forms and establish a library for coding the skeleton of the binary images

Abstract

In this paper, the methods for classifying points which are allowed to determining the types of two-dimensional vectors and can be used to code the skeleton of the binary image were analyzed. On the basis of , our research focused on the matrix which is t by t, in which there are only t elements are one. These elements can be called as point (pixel). An improved algorithm for classifying points is proposed. The result of the preprocessing point is a compressed library of point patterns.