Using Agent Technology for Recognizing Arabic Handwriting

Words as A phase of Document Analysis

Dr. Kamal M. Jambi

Computer Science Dept.King Abdul Aziz UniversityJeddah, Saudi Arabia

Abstract

This paper presents a module of a large research project funded by KACST. The work of this paper is based on Agent technology. This module is dealing with the process of Arabic character recognition. It gets from the main agent a block of text image and returns a file of characters contained in that image.  The process of recognition starts by identifying all structural features considered in this work with respect to Arabic character recognition. A word is segmented into its characters, then each character is recognized. The recognition is done by distributing features among special windows combined in a frame like a circle.

Keywords:

 Document Analysis, Arabic character recognition, Mobile Agents.