KIT | KIT-Bibliothek | Impressum
Open Access Logo
§
Volltext
DOI: 10.5445/IR/1000048047

On-line Recognition of Handwritten Mathematical Symbols

Thoma, Martin; Kilgour, Kevin; Stüker, Sebastian; Waibel, Alexander

Abstract:
This paper presents a classification system which uses the pen trajectory to classify handwritten symbols. Five preprocessing steps, one data multiplication algorithm, five features and five variants for multilayer Perceptron training were evaluated using 166898 recordings. The evaluation results of 21 experiments were used to create an optimized recognizer. This improvement was achieved by supervised layer-wise pretraining (SLP) and adding new features.


Zugehörige Institution(en) am KIT Institut für Anthropomatik und Robotik (IAR)
Publikationstyp Forschungsbericht
Jahr 2015
Sprache Deutsch
Identifikator ISSN: 2194-1629
URN: urn:nbn:de:swb:90-480476
KITopen ID: 1000048047
Verlag KIT, Karlsruhe
Umfang 8 S.
Serie KIT Scientific Working Papers ; 32
Schlagworte Recognition; machine learning; neural networks; symbols; multilayer perceptron
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft KITopen Landing Page