Robust Arabic Handwritten Word Recognition Technique Based on Wavelet Transform

Abstract

The Arabic Handwritten Word Recognition (AHWR) systems suffer from many types of challenges due to Arabic language characteristics. Nowadays, developing robust AHWR system is an attractive research topic due to the high demands in many commercial applications. A robust recognition system is proposed based on wavelet transform to improve the recognition rate for Arabic language recognition. The experimental work was done using IFN/ENIT dataset and LIBSVM library. The experimental results demonstrated that the proposed method provides significant improvements in recognition accuracy in compared to the methods, which based on DCT transform. The results show that the proposed methods succeeded in recognizing 78% of Arabic words. Moreover, the Principal Component Analysis is utilized in this research to improve the performance and to reduce the computational cost. The number of features was reduced to 124 features when using 5-level DWT with window of size $4X16$ pixels.