Development of Partial Face Recognition Framework

Author

Neo, H.F. ; Teo, C.C. ; Teoh, Andrew B J

Author_Institution

Fac. of Inf. Sci. & Technol., Multimedia Univ., Ayer Keroh, Malaysia

fYear

2010

fDate

7-10 Aug. 2010

Firstpage

142

Lastpage

146

Abstract

This paper presents a new framework of face recognition system that utilizes partial information available from front face where the users are willing and able to provide for access control. It is especially useful for people where part of their face is scarred and defect thus need to be covered. Hence, either top part/eye region (people is wearing veil or mask) or bottom part of face (people is wearing sunglasses) will be recognized respectively. The partial face data are tested with Principle Component Analysis (PCA), Non-negative matrix factorization (NMF), Local NMF (LNMF) and Spatially Confined NMF (SFNMF). The experiment results show that our proposed framework is able to achieve recognition rate of 95.17% with r = 80 by using SFNMF for bottom face region. On the other hand, eye region achieves 95.12% with r = 10 by using LNMF.

Keywords

authorisation; face recognition; matrix decomposition; principal component analysis; LNMF; PCA; SFNMF; access control; face recognition system; lcal NMF; nn-negative matrix factorization; partial face data; partial face recognition framework; principle component analysis; recognition rate; satially cnfined NMF; Approximation methods; Covariance matrix; Databases; Face; Face recognition; Feature extraction; Principal component analysis; local NMF; non-negative matrix factorization (NMF); partial face recognition; spatially confined NMF;

fLanguage

English

Publisher

ieee

Conference_Titel

Computer Graphics, Imaging and Visualization (CGIV), 2010 Seventh International Conference on

Conference_Location

Sydney, NSW

Print_ISBN

978-1-4244-7840-8

Type

conf

DOI

10.1109/CGIV.2010.29

Filename

5576207