A Bayesian model for recognizing handwritten mathematical expressions

作者：

Highlights：

• A recognizer for hand-drawn mathematical expressions.

• Relational context-free grammars structure hand-drawn mathematical input.

• Identifies multiple interpretations of the input based on output of lower-level classifiers.

• Bayesian probabilistic model for scoring and comparing recognized parse trees.

• Accuracy exceeds that of other academic recognizers.

摘要

Highlights•A recognizer for hand-drawn mathematical expressions.•Relational context-free grammars structure hand-drawn mathematical input.•Identifies multiple interpretations of the input based on output of lower-level classifiers.•Bayesian probabilistic model for scoring and comparing recognized parse trees.•Accuracy exceeds that of other academic recognizers.

论文关键词：Math recognition,On-line handwriting recognition,Bayesian networks,Stroke grouping,Symbol recognition,Relation classification

论文评审过程：Received 9 July 2014, Revised 31 October 2014, Accepted 18 February 2015, Available online 27 February 2015.

论文官网地址：https://doi.org/10.1016/j.patcog.2015.02.017