Zernike Moments and Neural Networks for Recognition of Isolated Arabic Characters
Mustapha Oujaoura, Rachid El Ayachi, Mohamed Fakir, Belaid Bouikhalene, Brahim Minaoui · viXra · 2012
The aim of this work is to present a system for recognizing isolated Arabic printed characters. This system goes through several stages: preprocessing, feature extraction and classification. Zernike moments, invariant moments and Walsh transformation are used to calculate the features. The classification is based on multilayer neural networks. A recognition rate of 98% is achieved by using Zernike moments.