Avaliação de conjuntos de características no reconhecimento de palavras manuscritas.

Abstract

This work presents a comparative evaluation of different feature sets used for handwritten word recognition. The main goal is to determine an optimum feature set to represent the handwritten names for the months of the year in Brazilian Portuguese language and to extend the conclusions obtained to other applications. For that purpose a baseline neural classifier was developed and used to determine the performance of the analysed feature sets. Three kinds of features are evaluated: perceptual, directional and topological. The evaluation shows that taken isolatedly. the perceptual feature set produces the best results for the lexicon used. These results can be improved combining the feature sets and the baseline' system with other classifier, in a hybrid approach, that obtained an average recognition rate of 90.4%.