Hierarchical attributed graph representation and recognition of handwritten chinese characters

摘要

This paper presents a new method of recognizing handwritten Chinese characters. A structural representation called hierarchical attributed graph representation (HAGR) is introduced to describe handwritten Chinese characters. The HAGR provides a simple and direct representation of handwritten Chinese characters. With HAGR, the recognition process becomes a simple task of graph matching. A cost function mapping a candidate to a model graph is introduced. This approach can tolerate the variations of HAGR which reflect the instabilities or variabilities of handwritten Chinese characters resulting from different writing styles. Several rules have been introduced to rearrange the order of the vertices of the graphs in order to avoid the combinatorial explosion in graph matching. In addition, the database of the character models is organized in a search-tree structure. For a candidate character, the search process to find a corresponding model character has been divided into a number of simple and local decisions at different levels of the tree. This considerably improves the efficiency and accuracy of the matching process.