Handwritten Arabic character recognition: which feature extraction method?

Recognition of Arabic handwriting characters is a difficult task due to similar appearance of some different characters. However, the selection of the method for feature extraction remains the most important step for achieving high recognition accuracy. The purpose of this paper is to compare the effectiveness of Discrete Cosine Transform and Discrete Wavelet transform to capture discriminative features of Arabic handwritten characters. A new database containing 5600 characters covering all shapes of Arabic handwriting characters has also developed for the purpose of the analysis. The coefficients of both techniques have been used for classification based on a Artificial Neural Network implementation. The results have been analysed and the finding have demonstrated that a Discrete Cosine Transform based feature extraction yields a superior recognition than its counterpart.