摘要
根据数学公式中字符或符号间空间关系特点,并针对目前用于数学公式字符空间关系判别的区域和质心方法所存在的不足,提出了基于字符凸壳和模糊识别的字符空间关系判别方法。首先,对数学公式中的字符或符号进行分类,对每一类运用字符凸壳判别其正上和正下关系,然后应用模糊识别的方法对常见关系进行识别。实验结果表明,运用该方法能明显提高字符空间关系判别的识别率,识别的正确率可达到93.5%。
This paper proposes a differentiating approach to the spatial relationship of characters based on symbol convex hull and fuzzy recognition theory in consideration of the particularities of such relationship among characters or symbols in mathematical expressions and in accordance with the insufficiency of the region and centroid method currently used for distinguishing the spatial relationships amongst characters in handwritten mathematical expressions. The characters or symbols in the mathematical expressions were firstly classified, and the above and below relationship was discriminated for each class with symbol convex hull,then the familiar relationship was recognized with fuzzy method. The experimental results on the paper demonstrate that the recognition rate of the spatial relationship of characters was greatly improved by the method with the precision rate of 93.5%.
出处
《计算机应用与软件》
CSCD
北大核心
2008年第5期208-209,254,共3页
Computer Applications and Software
关键词
数学公式识别
关系识别
凸壳
模糊识别
Recognition of mathematical expressions Relationship recognition Convex hull Fuzzy recognition
