New methods and databases for hand biometrics

Other Title(s)

طرق و قواعد بيانات جديدة لتمييز الأشخاص اعتمادا على خصائص اليد البشرية

Dissertant

al-Butush, Iman Salamah Salim

Thesis advisor

al-Hasanat, Ahmad Bashir

Comitee Members

al-Mahadin, Bassam Muhammad Salim
al-Abbadi, Muhammad Ali Husayn
al-Kasasibah, Malik Zakariyya

University

Mutah University

Faculty

Information Technology College

Department

Computer Science Department

University Country

Jordan

Degree

Master

Degree Date

2016

English Abstract

A biometric identification/verification system verifies user identities by comparing some sort of behavioral or physiological trait of the user to a stored sample in database.

Several states of the art were reviewed and the most commonly used geometrical features were identified.

This thesis introduces new hand images databases with total 6500 hand images.

The first one is created by using a mobile phone camera of modest quality, and called mobile hand images database (MOHI), the second one created by using a webcam with low quality images, and called webcam hand images database (WEHI).

The last one, called victory sign hand images database (VSHI).

This database contain the victory signs of 50 different persons captured using a mobile phone camera.

The major goal, of those databases, is to build upon person’s identification/verification using hand biometrics.

Hand shape biometric system consists of a sequence of several major steps, including: Image Acquisition, Hand Segmentation, Feature Extraction, Training/Testing and storing the hand templates for identification.

Image acquisition is done by creating the databases.

Segmentation process includes investigating different methods such as Otsu's method, the k-means clustering based on the color information (RGB color model), and hand segmentation based on color information using Artificial Neural Network (ANN).

Chain Code was used to extract 20 hand features from MOHI and WEHI, with accuracy greater than 85%.

Moreover; this thesis introduces a new approach for identifying persons (e.g.

terrorist) from their geometric features of victory sign alone, as this sign might be the only information available about such persons.

For this purpose, VSHI database created and investigated some simple approaches to identify persons from those images.

The results of this database were encouraging for most of the recorded persons; with accuracy range from 40% to 93%, depending on the features, distance metric and the constant K used.

Finally analytical and comparative study in segmentation and feature extraction introduced over those databases

Main Subjects

Information Technology and Computer Science

No. of Pages

62

Table of Contents

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-731647