Face recognition under partial occlusions based on Eigenvectors of face feature blocks

Other Title(s)

التعرف على الوجوه بوجود حجب جزئي اعتمادا على المتجهات الذاتية (Eigenvectors)‎ لمجتزءات خصائص الوجه

Dissertant

Abu Zayd, Marwan Saleh Hassan

Thesis advisor

al-Attar, Ashraf Muhammad

University

Islamic University

Faculty

Faculty of Information Technology

Department

Information Technology

University Country

Palestine (Gaza Strip)

Degree

Master

Degree Date

2017

English Abstract

In the presence of partial face occlusion, existing face recognition methods typically face the challenge of achieving both high accuracy and minimum computation time.

High recognition accuracy requires extensive computation which significantly decreases time performance, and vice versa.

Furthermore, occlusion significantly degrades the performance of face recognition systems.

In this research, we propose a new method to achieve both high recognition accuracy in the presence of partial occlusion.

Our method is based on extracting separate face feature blocks (FFB), and calculating the eigenvectors for each block.

Eigenvectors of query image FFBs are matched with eigenvectors of database FFBs using Euclidian distance as similarity measure.

Our proposed method has the advantage of increasing the retrieval accuracy From the experimental results, it is evident that our system performs significantly better compared with other existing methods.

Evaluation of our method shows a 98% face recognition accuracy under partial occlusion on scarf case and a 93.5% on sunglasses case.

This accuracy level is better accuracy with other existing methods.

The experimental evaluation of the method is based on AR dataset.

Main Subjects

Information Technology and Computer Science

No. of Pages

60

Table of Contents

• APA
• MLA
• AMA

Language

English

Data Type

Arab Theses

Record ID

BIM-905693