Morphological skeleton traversal algorithm for person recognition based on handwritten stroke analysis

Abstract

Traversing in the sense and direction of the writing the morphological skeleton of a grapheme, is a key element in the construction of a series of recent descriptors for person recognition based on the analysis of strokes of handwriting. In this paper, the traversal of the skeleton as a problem of traversal in graphs is formulated and an algorithm based on the Tables of Priority and Backtracking and Bound is presented. A wide base of data of more than 4500 skeletons of the grapheme “e” is constructed to carry out the experiments; it is shown that the theoretical formulation of the problem and the employed algorithmic strategy allows encompassing successfully complicated things of skeletons that possess crosses.