手写印刷体汉字识别的细化算法研究

The Study on Thinning Algorithm of Handwritten Printed Chinese Character Recognition

细化是汉字识别预处理中最重要的技术之一。它可在保持原汉字大部分信息的基础之上，极大地压缩汉字的存储量，有利于后续的特征抽取与识别。在比较了几种不同细化方法的优劣之后，提出了一种改进的细化算法（以下简称Ｉ－ＦＡ算法），并进行了计算机仿真。实验结果表明，Ｉ－ＦＡ无论在速度或细化质量上都较文中的另几种算法有较明显的提高。可用于进一步的文字识别软件的研制中。

Thinning is one of the important techniques in the preprocessing of Chinese Character Recognition (CCR).Several methods of thinning are compared in this paper and an improved algorithm is proposed.Simulation results show that our algorithm has considerable improvement in speed and quality over the existing methods.It can be further used in the study of CCR software.