摘要
目前印刷体数学公式识别系统的输出还存在着误识结果,进行必要的后处理是提高识别率的重要手段。介绍了一种将印刷体数学公式识别结果与公式的语义知识相结合的方法,对其误识结果进行系统的分析,给出了若干条共有的规则及基准转移等方法,进行综合纠错的后处理,从而进一步完善印刷体数学公式识别系统。实验结果表明,该方法能够有效地提高系统识别结果的正确率。
At present, the output of the printed mathematical expressions recognition system still exist inaccurate result and suitable post-processing is the important way in order to improve recognition rate. Firstly one kind of method which unified the printed mathe- matical formula structure analysis and semantics knowledge is introduced and then the wrong results are analyzed systematically. The methods of several rules in common and baseline shift are given so that the post-processing based on synthetic error correction is carried on. Thus printed mathematical formula recognition system gets consummate. The experiment indicates that the accuracy of the system is enhanced effectively.
出处
《计算机工程与设计》
CSCD
北大核心
2007年第20期5039-5041,5044,共4页
Computer Engineering and Design
基金
国家自然科学基金项目(60772073)
关键词
公式识别
结构分析
后处理
规则
基准线转移
mathematical expressions recognition
structural analysis
post-processing
rules
baseline shift
