Features for HMM-Based Arabic Handwritten Word Recognition Systems

Abstract

HMM-based systems need observation sequences as input. Such observa- tions consist of discrete values or vectors extracted from word images or text lines. We explore in this chapter various types of features which are popular for Arabic cursive handwriting recognition. Such features are statistical, based on pixel distri- butions or local directions. Others are structural, based on the presence of loops, ascenders or descenders.We show how these features can be efficient within HMM- based systems based on sliding-windows or grapheme segmentation.