一种新的手写汉字生成方法

Generation of a New Kind of Handwritten Chinese Character

改进的汉字统计结构模型可生成给定风格下的手写汉字。汉字被分为三个层次:笔划、部首和单字,我们首先训练样本,基于主成分分析和核主成分分析,分别建立三个层次的概率分布;然后测试样本,最后生成了与测试样本同一风格的汉字。使用HCL2000汉字数据库进行实验,实验结果验证了提出模型的有效性。

An improved character structure model to generate handwritten Chinese characters under given style is proposed. Chinese characters are decomposed by three levels：stroke, radical and single character. Firstly, we train samples and respectively build distribu- tions of three levels based on principal component analysis and kernel principal component analysis ： then we test samples ： and finally, the characters under the same style as test samples are generated. This paper experiments on Chinese characters HCL2000 database, and the results show that the proposed model is effective.