Computing Component Alignment Normalization Degree in Digital Ink Chinese Character from International Students Based on Information Database
Xijin Sun, Xiwen Zhang · 2021
There are many irregularities of components in the handwritten Chinese character by junior international students. Among them, the problem of component alignment is more prominent. For digital ink Chinese characters (DICCs), this paper proposes a difference computing approach based on the component alignment information database of reference Chinese characters (RCCs). The component alignment database stores information related to various types of component alignments, and facilitates automatically extracting handwritten Chinese character components by reading the structural information in the database. Component alignment computing is based on the barycenter approach, which takes full account of the structure type and component layout rule of Chinese characters, supplemented by component boundary alignment computing, and the average value of each alignment data is normalized. Using the component alignment data of the RCC as a standard, the difference between the component alignment of the DICC and the RCC is computed to obtain the alignment difference. The experimental data includes three kinds of Chinese character structures, which are suitable for computing ten kinds of structures of Chinese characters. Experimental results show that the computed component alignment difference is consistent with human subjective visual perception. The quantitative value of the alignment difference can provide correction data, which is convenient for the visual correction of subsequent work and the teaching of component writing. It can also perform statistical analysis and rule discovery, and provide data support for the study of Chinese character acquisition.