摘要
提出了一种实时高效的手写人群速记符三级识别策略,在以声符和韵符为识别基元的基础上,将基于速记符知识的特征规则法和基于速记符生成模型的模板匹配法有效地结合起来,对速记符的297 年基本音符进行了识别实验结果表明,系统达到了较高的性能指标本方案已在286 和486 等微机上实现,经训练识别率可达97%
A 3 stage recognition strategy to chinese handwritten shorthand is pre sented. Using initial consonant notation and rhyme notation as recognition primitive ele ment, this stategy combines the feature rule approach based on knowledge of shorthand with the pattern matching approach based on the shorthand phonetic notation model during the recognition of 297 pinyin notations. Expermental results show that the system reaches fairly good performace. This new approach has been implemented on the models 286 and 486 microcomputer, and its accurate recognition rate can reach over 97%.m ent, this stategy com bines the feature rule approach based on know ledge of shorthandw ith the pattern m atching approach based on the shorthand phonetic notation m odelduring the recognition of297 pinyin notations. Experm entalresults show that the systemreaches fairly good perform ace. This new approach has been im plem ented on the m odels286 and 486 m icrocom puter, and its accurate recognition rate can reach over 97% .
基金
山东省科委资助