Arabic Bank Cheque Words Recognition Using Gabor Features

Abstract

Arabic cheque processing is one of the important applications of handwriting recognition. The recognition of Arabic Cheque bank is still awaiting lots of work in its constituent stages, which include pre-processing, feature extraction and classification. Several feature extraction methods used to recognize handwritten digits and words. The stroke direction is one important feature of Arabic handwriting which Gabor filter proved its ability to detect this local structural feature. On the other hand, investigating different classifiers can improve the recognition accuracy. In this paper, Gabor features are investigated with ELM and SMO classifiers. Two Arabic Cheque datasets, AHDB and CENPARMI, are used for evaluation. The results from Gabor features with SMO classifier outperform previous studies.