A Neural Network Model for Online Handwritten Mathematical Symbol Recognition

Abstract

This paper proposes a new handwritten mathematical symbol recognition system that is flexible enough to let the users write the symbols in their own ways. They do not have to learn a completely new way of writing symbols. The proposed approach involves two main stages: online and offline. During the online stage, the input is classified into one of the four groups. During the offline stage, the new neural network, called Hausdorff ARTMAP, which is specifically designed for solving two dimensional binary pattern recognition problems is used to identify the symbols. The proposed model is tested in a writer independent mode using the researcher’s own collected database. The result obtained is very encouraging.