Observations and Issues on Symbol Recognition in Online Mathematical Expression Recognition(Character and document processing)

概要

論文の詳細を見る
Online mathematical expression recognition is mainly composed of two major phases:(1) symbol recognition phase and (2) structural analysis phase. Though it is known that the former phase is generally easier than the latter one, the effort of reducing grouping/recognition errors is still important because not only wrong symbol grouping/recognition results often give bad effect to the structural analysis phase, but also it is useful to reduce search space of the latter phase. In this paper, we first observe the characteristics of symbol classes for on-line mathematical expression recognition compared with ones for general handwriting recognition, and then devise a simple recognition method for the mathematical symbols which are used in the mathematical expressions on the Korean high school mathematics textbook. The process can be applied to the recognition-based symbol grouping. Additionally, several geometrical constraints are designed to reject invalid symbol groupings effectively.
2006-11-17