Understanding Mathematics: A System for the Recognition of On-Line Handwritten Mathematical Expressions [article]

Ernesto Tapia, Universitätsbibliothek Der FU Berlin, Universitätsbibliothek Der FU Berlin
2005
This work presents a system for the recognition of on-line handwritten mathematical formulas. The system consists of two main stages: Classification of isolated on-line handwritten symbols and the analysis of spatial relationships among them. We propose a system for the recognition of isolated on-line handwritten characters which is based on support vector classification. We also propose a suitable representation for strokes and symbols which is used to improve the classification rates of the
more » ... assifier. Our experiments show that our classifier achieved better classification rates in comparison to other popular classification techniques. This could be accomplished by extensive preprocessing of the data and by parameter selection for the support vector classification. We propose a new structural analysis method for the recognition of on-line handwritten mathematical expressions based on a minimum spanning tree construction and symbol dominance. Our method addresses important layout problems frequently encountered in on-line handwritten formula-recognition systems. Our method also aims to handle input as naturally as possible, i.e. using the usual mathematical conventions, without restrictions in the order the symbols are written. Our method handles symbols with non-standard layout, like \sideset{^{*}_{*}}{^{*}_{*}}\prod, as well as tabular layouts, e.g. matrices. Our system for the recognition of on- line handwritten mathematical expressions is used in the Electronic Chalkboard (E-Chalk), a multimedia system for distance-teaching.
doi:10.17169/refubium-16590 fatcat:b4fevf76vndgti5v7hcukq4dpe