联机手写中文速记符的自动识别

Automatic Recognition of On-Line Handwritten Chinese Shorthand

手写中文速记符的自动识别是文字类识别中的一个比较特殊的课题 .实现了一个联机手写中文速记符识别系统 HCSRS( handwritten Chinese shorthand recognition system) ,并给出了对速记符中的基本音符、独立略符、连笔符的识别实验结果 .该识别系统采用了以声符、韵符为基元的结构识别策略 .为了实现基元的有效切分 ,提出了一种基于切分 -分析 -交互结构及其相应知识库的自调整切分算法 STSA ( self- tuningsegmentation algorithm) ,从而大大改进了系统的识别性能 .理论分析和测试结果表明 ,利用计算机来实现手写中文速记符到书面汉字的同步转换是可行的 .

Automatic recognition of handwritten Chinese shorthand is a special issue of cha racter recognition subject. There is little literature on this subject. This pap er implements an on-line recognition system for HCSRS (handwritten Chinese shor thand recognition system) and provide experimental results for the recognition o f basic phonetic shorthands, shortforms, and linking shorthands. In this system, a structural recognition strategy based on the primitives of initial consonant symbol and rhyme symbol is applied. For efficient segmentation of primitives, a novel STSA (self-tuning segmentation algorithm) based on the segmentation-anal ysis-interaction configuration and corresponding knowledge sources is presented in this paper, which greatly improves the performance of the system. Theoretica l analysis and experimental results show that it is feasible to realize the simu ltaneous transformation from handwritten Chinese shorthands to written Chinese c haracters through computer.