An Efficient Preprocessing Framework for Arabic Handwriting Recognition System

Abstract

Image preprocessing has assumed an essential part in Arabic Handwriting Recognition System (AHRS). It has several stages which affect the accuracy of the recognition system. An efficient preprocessing framework for Arabic handwriting recognition system has been proposed in this paper. Since preprocessing stage in AHRS is imperative to reduce the dimensionality of image to remove the undesirable information then increase the processing speed of the AHRS. Besides, it provide correction process for the dataset elements to make it work probably through several processes that will discuss in this paper. Exploratory results with artificial and real life images shows that, the proposed method gives an efficient results that help to get a high recognition rate in the AHRS.