Structural and Statistical Feature Extraction Methodology for the Recognition of Handwritten Arabic Words

Abstract

Knowledge concerning the topography of Arabic letters, as well as the structural characteristics between background regions and character components is investigated as a novel approach for Arabic recognition. The suggested feature extraction method reduces the classifier input data to only the most significant and essential. First, connected components consisting of more than one character are segmented into characters. Secondly, the primitives are extracted according to the knowledge of character structures and some statistical characteristics. Finally a hybrid model based on the combination of support vector machines (SVM) classifier and particle swarm optimization (PSO) is used to evaluate the performance of features extracted.