Face verification using modified matched spatial filter

Martin Leonard Tangel; Muhammad Rahmat Widyanto

Face verification using modified matched spatial filter

Martin Leonard Tangel, Muhammad Rahmat Widyanto

Department of Computer Science

Research output: Contribution to conference › Paper › peer-review

Abstract

To improve recognition capability of face verification system, Modified Matched Spatial (MMS) Filter method is proposed in this research. MMS Filter can cope with facial expression problem due to its capability to construct sharp peak value of correlation output. The peak value itself is obtained from power spectrum that represents particular frequency. In experiment using facial expression database from Advanced Multimedia Processing Laboratory Carnegie Mellon University, MMS Filter has outperformed Matched Spatial Filter by obtaining 100% for its accuracy under ideal condition. The development of face verification system using MMS filter will produce a robust face verification system from various facial expression.

Original language	English
Publication status	Published - 2009
Event	International Workshop on Advanced Computational Intelligence and Intelligent Informatics, IWACIII 2009 - Tokyo, Japan Duration: 7 Nov 2009 → 7 Nov 2009

Conference

Conference	International Workshop on Advanced Computational Intelligence and Intelligent Informatics, IWACIII 2009
Country/Territory	Japan
City	Tokyo
Period	7/11/09 → 7/11/09

Keywords

Correlation filter
Face verification
Matched spatial filter

Cite this

@conference{3450d9336aa64bb8b834d243acaf2373,

title = "Face verification using modified matched spatial filter",

abstract = "To improve recognition capability of face verification system, Modified Matched Spatial (MMS) Filter method is proposed in this research. MMS Filter can cope with facial expression problem due to its capability to construct sharp peak value of correlation output. The peak value itself is obtained from power spectrum that represents particular frequency. In experiment using facial expression database from Advanced Multimedia Processing Laboratory Carnegie Mellon University, MMS Filter has outperformed Matched Spatial Filter by obtaining 100% for its accuracy under ideal condition. The development of face verification system using MMS filter will produce a robust face verification system from various facial expression.",

keywords = "Correlation filter, Face verification, Matched spatial filter",

author = "Tangel, {Martin Leonard} and Widyanto, {Muhammad Rahmat}",

year = "2009",

language = "English",

note = "International Workshop on Advanced Computational Intelligence and Intelligent Informatics, IWACIII 2009 ; Conference date: 07-11-2009 Through 07-11-2009",

}

TY - CONF

T1 - Face verification using modified matched spatial filter

AU - Tangel, Martin Leonard

AU - Widyanto, Muhammad Rahmat

PY - 2009

Y1 - 2009

N2 - To improve recognition capability of face verification system, Modified Matched Spatial (MMS) Filter method is proposed in this research. MMS Filter can cope with facial expression problem due to its capability to construct sharp peak value of correlation output. The peak value itself is obtained from power spectrum that represents particular frequency. In experiment using facial expression database from Advanced Multimedia Processing Laboratory Carnegie Mellon University, MMS Filter has outperformed Matched Spatial Filter by obtaining 100% for its accuracy under ideal condition. The development of face verification system using MMS filter will produce a robust face verification system from various facial expression.

AB - To improve recognition capability of face verification system, Modified Matched Spatial (MMS) Filter method is proposed in this research. MMS Filter can cope with facial expression problem due to its capability to construct sharp peak value of correlation output. The peak value itself is obtained from power spectrum that represents particular frequency. In experiment using facial expression database from Advanced Multimedia Processing Laboratory Carnegie Mellon University, MMS Filter has outperformed Matched Spatial Filter by obtaining 100% for its accuracy under ideal condition. The development of face verification system using MMS filter will produce a robust face verification system from various facial expression.

KW - Correlation filter

KW - Face verification

KW - Matched spatial filter

UR - http://www.scopus.com/inward/record.url?scp=79960409772&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:79960409772

T2 - International Workshop on Advanced Computational Intelligence and Intelligent Informatics, IWACIII 2009

Y2 - 7 November 2009 through 7 November 2009

ER -

Face verification using modified matched spatial filter

Abstract

Conference

Keywords

Other files and links

Fingerprint

Cite this