Printed Arabic Characters Recognition Using Combined Features and CNN classifier
Lallouani Bouchakour, Fariza Meziani, Houda Latrache, Khadija Ghribi, Mustapha Yahiaoui · 2021
In this paper we investigate the optical characters recognition for Arabic language (AOCR). This system is considered as a challenging research topic due to richness and difficulties of Arabic writing. The OCR system incorporates three main stages that are segmentation, feature extraction and recognition. In this work, we have proposed a new method to recognize printed Arabic characters. This method is based on the combined features extraction, which are the densities of black pixels, invariant moments of Hu and Gabor features and the Convolution Neural Network CNN classifier. Experiments are conducted on the Printed Arabic Text set PAT-A01.The result show that the features combination enhances the recognition accuracy rate.