Details

Autor(en) / Beteiligte

• Garain, U.
• Chaudhuri, B.B.
• Ghosh, R.P.

Titel

A multiple-classifier system for recognition of printed mathematical symbols

Ist Teil von

• Proceedings of the 17th International Conference on Pattern Recognition, 2004. ICPR 2004, 2004, Vol.1, p.380-383 Vol.1

Ort / Verlag

IEEE

Erscheinungsjahr

2004

Link zum Volltext

• IEEE_Xplore

Quelle

IEEE Electronic Library (IEL)

Beschreibungen/Notizen

• This paper deals with recognition of printed mathematical symbols. A group of classifiers arranged hierarchically is used to achieve robust recognition of the large number of symbols appearing in expressions. The classifier used at the top level employs stroke-based classification technique to recognize some of the frequently occurring symbols. The second level uses three classifiers to recognize the rest of the expression symbols. Different combination techniques have been attempted to integrate the second level classifiers to achieve high recognition accuracy. Experiment shows that the proposed approach is quite robust for recognition of a large number of symbols appearing in various expressions.