数学公式图像处理综述

Survey of Mathematical Expression Image Processing

数学公式广泛存在于各类文献之中,因此数学公式的自动定位、识别、分析和理解是文档图像处理必须研究的问题。但是数学公式和普通文本存在很大区别,所以数学公式的识别、分析远比文字段落的识别困难。本文回顾了数学公式图像处理的研究历程,提出了公式处理的模型,总结比较了公式定位、公式识别、公式分析以及性能评估等方面的主要处理方法,并展望了未来的研究方向。

Mathematical expressions exist in all kinds of documents . Thus , the automatic location , recognition, analysis and understanding of mathematical expressions are important to document image processing. However, because of the many differences between mathematical expressions and normal texts, expression processing is far more difficult than ordinary text recognition. This paper reviews the history of mathematical expression image processing, presents the modules of expression processing, surveys the primary processing methods of expression extraction, expression recognition, expression analysis and performance evaluation, and gives the perspective of future research directions.