Features extraction using thermal face image recognition

Other Title(s)

استخراج الخصائص باستخدام التعرف على الصورة الحرارية للوجه

Dissertant

Shnasi, Thura Muataz

Thesis advisor

al-Mashaykhi, Akram M. Uthman

Comitee Members

Riadh, Mayy Haikil
Kanan, Ghassan

University

Amman Arab University

Faculty

Collage of Computer Sciences and Informatics

Department

Department of Computer Science

University Country

Jordan

Degree

Master

Degree Date

2014

English Abstract

Face recognition, using thermal images, has become an area of growing interest due to the increased demand of security applications in many organizations such as airports, banks …etc.

Thermal face recognition is a robust system that is employed under all lighting conditions including total darkness also if the person is wearing a disguise, fake nose or make-up.

The main problem in thermal face recognition is how to extract the features that distinguish one person from another.

In this work, a model is proposed for the extraction of physiological features to identify the person based under different expressions and illuminations.

The model is comprised of four logically sequential stages aimed at verifying the results then their analysis.

In the processing of this model, IRIS (Imaging Robotics and Intelligent System) database is used.

First, many enhancement filters are used in the pre-processing module to enhance the images, morphological operations and noise removing filters are presented.

Secondly, blob detection is employed in the detection module to detect the features blobs regions.

These blobs represent the region of eyes, mouth, and nose.

In the third stage of the features extraction module, a set of geometrical features such as distances, slopes, and center points between the extracted blobs are computed.

In the fourth stage of the classification module, the proposed model is employed.

Good results of features extraction and image classification are obtained.

The results of the proposed model during the experimental treatment of different individuals and number of images varied between 55% and 85%.

From the analysis of results, the relative variation for some images is attributed to the poor quality of images used in the experimental treatment.

Finally, the thesis presents some conclusions and suggestions for future work to develop the proposed model further

Main Subjects

Electronic engineering

Topics

• Digital electronics

No. of Pages

81

Table of Contents

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-561846