Header Information

NPRP 09 - 864 - 1 - 128
NPRP 03
Qatar University
Award Closed
01 Dec 2010
Dr. Somaya Al-Maadeed
3 Year(s)
01 Sep 2014
Resubmission of [NPRP 08 - 313 - 1 - 059]
Intelligent Document Management System for Automatic Writer Identification

Project Summary
The objective of this project is to develop an intelligent document management system that can be used for Arabic forensic writer identification in the Qatar Government’s forensic labs. This will be achieved by developing new automatic pre-processing and classification procedures to handwritten documents, in order to use them in forensic writer identification. This will enable writer identification, based on the stored documents and other data in the database. That is, any text containing common words in Arabic handwriting may be used to establish the identity of the writer. To achieve these goals the project will have four phases as follows: 1- Build a system for documents background image removal. Removing the background while still enhancing the quality of the data is essential for feature extraction phase, as this increase the recognition rate. A fuzzy based thresholding technique will be implemented on the documents to remove the background. 2- Build an intelligent document management system to help forensic scientists in document examination and handwriting identification. 3- Develop a novel feature extraction system to extract distinctive and most discriminating features from handwritten documents. 4- Develop an intelligent writer identification system. Both k-nearest/neural conceptual approach and Kohonen network classifiers will be investigated in order to generate codebooks of graphemes for use to recognize the writers based on their handwriting.
The Ministry of Interior in Qatar raised their need for a program to help them in the identification of writers of unknown documents. This project aims to develop a solution to satisfy these requirements. Another problem of writer identification arises frequently in the court of justice in Qatar, where one must come to a conclusion about the authenticity of a document (e.g. a will). It also arises in Qatari banks for signature verification, or in some institutes which analyze texts of former authors, and are interested in the genetics of these texts, as for example the identification of the various writers who took part in the drafting of a manuscript or who made corrections. We will develop new methods for Arabic handwriting image enhancement using High Boost filtering with a view to lead to efficient image binarization. As a starting point, we will use methods employed by current fingerprint methods. We will investigate a number of approaches using simple and more complex classifiers such as k-NN, conceptual method and Kohonen neural networks and analyze their performances in terms of classification accuracy and speed. Benefits are also expected to accrue for the PIs in terms of collegiality and productivity, and encouraging of technology makers to look to Qatar as a potential hotbed for education and research.
Document Management; Writer Identification; Image processing; Pattern Recognition
Applied research
1. Natural Sciences
1.2 Computer and Information Sciences
Computer Sciences
Yes
No

Institution
Qatar University
Qatar
Performance Site
Qatar University
Qatar
Submitting Institution
Northumbria University
United Kingdom
Collaborative Institution
Northumbria University
United Kingdom
Performance Site

Personnel
Lead PI
Dr. Somaya Al-Maadeed
Qatar University
Co-Lead PI
Dr. Somaya Al-Maadeed
Qatar University
Co-PI
Prof. Ali Jaoua
Qatar University
Co-PI
Prof. Ahmed Bouridane
Northumbria University
Co-PI
Prof. Jihad Mohamad ALJA'AM
Qatar University

Outputs/Outcomes
Conference Paper
ICFHR2012 Competition on Writer Identi?cation -
Abdelaali Hassaine, Somaya Al Maadeed
DOI:10.1109/ICFHR.2012.218
Conference Paper
QUWI: An Arabic and English Handwriting Dataset
Somaya Al Maadeed, Wael Ayouby, Abdelaali Hassaine, Jihad Mohamad Aljaam
DOI:10.1109/ICFHR.2012.256
Conference Paper
Feature-based method for offline writer identification
Somaya Al-Maadeed, Abdelaali Hassaine
DOI:10.5339/qfarf.2012.CSP29
Conference Paper
Detecting forgeries and disguised signatures in online signature verification
Abdelaali Hassaine, Somaya Al-Maadeed
DOI:10.5339/qfarf.2012.CSP30
Conference Paper
Multi-modal biometric authentication system using face and online signature fusion
Youssef Elmir, Somaya Al-Maadeed, Abbes Amira, Abdelaali Hassaine
DOI:10.5339/qfarf.2012.CSP32
Online Paper
Text-Dependent Writer Identification for Arabic Handwriting
Somaya Al-Maadeed
DOI:10.1155/2012/794106
Book
Neural Information Processing Lecture Notes in Computer Science Volume 7667, 2012, pp 576-583
Youssef Elmir, Somaya Al-Maadeed, Abbes Amira, Abdelaali Hassaïne
ISBN:9783642344992
Conference Paper
the ICDAR2011- Arabic Writer Identification Contest
Abdelâali Hassaïne
DOI:10.1109/ICDAR.2011.292
Conference Paper
Skeleton-Based Codebooks for Writer Identification
Abdelâali Hassaïne
DOI:QNRFOLDPUB114
Conference Paper
The ICFHR 2012 Writer Identi?cation Contest, Challenge 2: Arabic Scripts
Abdelaali Hassaine
DOI:10.1109/ICFHR.2012.218
Conference Paper
A New Handwriting Dataset for Automatic Writer Identification Systems
Abdelaali Hassaine
DOI:QNRFOLDPUB186
Journal Paper
ARABIC SIGNATURE VERIFICATION DATASET
SOMAYA AL-MAADEED, WAEL AYOUBY, ABDELAALI HASSAINE, AMNA ALMEJALI, AMAL AL-YAZEEDI, and ROUDHA AL-ATIYA
ISSN:18120857
Online Paper
Automatic prediction of age, gender, and nationality in offline handwriting
Somaya Al Maadeed-author-first, Abdelaali Hassaine-author-additional
DOI:10.1186/1687-5281-2014-10