Pose-invariant face recognition

Pao Yue-kong Library Electronic Theses Database

Pose-invariant face recognition


Author: Wang, Jiannan
Title: Pose-invariant face recognition
Degree: M.Sc.
Year: 2014
Subject: Face perception -- Data processing.
Human face recognition (Computer science)
Hong Kong Polytechnic University -- Dissertations
Department: Faculty of Engineering
Pages: iv, 57 leaves : illustrations ; 30 cm
Language: English
InnoPac Record: http://library.polyu.edu.hk/record=b2680823
URI: http://theses.lib.polyu.edu.hk/handle/200/8135
Abstract: Face recognition has been widely used in security systems. The accuracy of these systems is very important for most of the applications. However, the recognition performance of most face-recognition algorithms is significantly influenced by the variations in facial appearances caused by pose variations. One of the best solutions to this problem is to generate the virtual frontal-view image from a given non-frontal-view face image under a specific pose. In this thesis, we have investigated different techniques for pose-invariant face recognition. In particular, we have employed a linear-regression-based method to handle the pose problem. In this approach, we assume that there is a linear relationship between a frontal-view face image and its corresponding non-frontal-view counterpart. We can estimate a linear mapping for the whole face and the local face regions; thus, we employ the Globally Linear Regression and the Locally Linear Regression methods, respectively, for frontal-view face reconstruction. We have compared the pose-invariant face-recognition method with Principal Component Analysis, i.e. the Eigenface method, and the FisherFace method. All of these different methods have been evaluated based on the CMU PIE database, and our experiment results show that the pose-invariant method can achieve a much better performance than the other methods when the pose variation is large.

Files in this item

Files Size Format
b26808237.pdf 5.253Mb PDF
Copyright Undertaking
As a bona fide Library user, I declare that:
  1. I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
  2. I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
  3. I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.


Quick Search


More Information