Recognition of standard arabic style

Dissertant

Fluah, Layth A.

University

University of Technology

Faculty

-

Department

Department of Control and Systems Engineering

University Country

Iraq

Degree

Master

Degree Date

2000

English Abstract

Recognition of Arabic style is a difficult task in the area of character recognition because Arabic is a cursive language and differs in structure from other languages. Relatively few studies have been devoted to cursive Arabic hand-written recognition as this field has many challenging problems due to the variations in writing style and the lack of dynamic information related to the shape of the character. One of these challenging problems is the segmentation of words that have overlapping in characters, segmented these overlapped characters is difficult as the segmentation point within the character itself and cannot be dedicate, in these thesis we found a method to overcome this problem in the NASKH style and by using this method we could segmented words written in NASKH style have overlap in characters in efficient way.

We use Artificial Neural Network (ANN) techniques to implement the recognition of characters, Perception Model has been used.

Finally the output of the proposed system is the ASCII code of recognized characters.

Main Subjects

Information Technology and Computer Science

Topics

• Writing
• Data processing
• Computational linguistics
• Arabic language

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-306209