LBP and SIFT based facial expression recognition

Omer Sumer; Ece Olcay Gunes

doi:10.1117/12.2181505

LBP and SIFT based facial expression recognition

Omer Sumer, Ece Olcay Gunes

Department of Electronics and Communication Engineering

Istanbul Technical University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

1 Citation (Scopus)

Abstract

This study compares the performance of local binary patterns (LBP) and scale invariant feature transform (SIFT) with support vector machines (SVM) in automatic classification of discrete facial expressions. Facial expression recognition is a multiclass classification problem and seven classes; happiness, anger, sadness, disgust, surprise, fear and comtempt are classified. Using SIFT feature vectors and linear SVM, 93.1% mean accuracy is acquired on CK+ database. On the other hand, the performance of LBP-based classifier with linear SVM is reported on SFEW using strictly person independent (SPI) protocol. Seven-class mean accuracy on SFEW is 59.76%. Experiments on both databases showed that LBP features can be used in a fairly descriptive way if a good localization of facial points and partitioning strategy are followed.

Original language	English
Title of host publication	Seventh International Conference on Machine Vision, ICMV 2014
Editors	Branislav Vuksanovic, Jianhong Zhou, Antanas Verikas, Petia Radeva
Publisher	SPIE
ISBN (Electronic)	9781628415605
DOIs	https://doi.org/10.1117/12.2181505
Publication status	Published - 2015
Event	7th International Conference on Machine Vision, ICMV 2014 - Milan, Italy Duration: 19 Nov 2014 → 21 Nov 2014

Publication series

Name	Proceedings of SPIE - The International Society for Optical Engineering
Volume	9445
ISSN (Print)	0277-786X
ISSN (Electronic)	1996-756X

Conference

Conference	7th International Conference on Machine Vision, ICMV 2014
Country/Territory	Italy
City	Milan
Period	19/11/14 → 21/11/14

Bibliographical note

Publisher Copyright:
© 2015 SPIE.

Keywords

emotion classification
Facial Expression Recognition (FER)
LBP
SIFT
SVM

Access to Document

10.1117/12.2181505

Cite this

@inproceedings{c41bb353aa3d4291be424b3520cdbb37,

title = "LBP and SIFT based facial expression recognition",

abstract = "This study compares the performance of local binary patterns (LBP) and scale invariant feature transform (SIFT) with support vector machines (SVM) in automatic classification of discrete facial expressions. Facial expression recognition is a multiclass classification problem and seven classes; happiness, anger, sadness, disgust, surprise, fear and comtempt are classified. Using SIFT feature vectors and linear SVM, 93.1% mean accuracy is acquired on CK+ database. On the other hand, the performance of LBP-based classifier with linear SVM is reported on SFEW using strictly person independent (SPI) protocol. Seven-class mean accuracy on SFEW is 59.76%. Experiments on both databases showed that LBP features can be used in a fairly descriptive way if a good localization of facial points and partitioning strategy are followed.",

keywords = "emotion classification, Facial Expression Recognition (FER), LBP, SIFT, SVM",

author = "Omer Sumer and Gunes, {Ece Olcay}",

note = "Publisher Copyright: {\textcopyright} 2015 SPIE.; 7th International Conference on Machine Vision, ICMV 2014 ; Conference date: 19-11-2014 Through 21-11-2014",

year = "2015",

doi = "10.1117/12.2181505",

language = "English",

series = "Proceedings of SPIE - The International Society for Optical Engineering",

publisher = "SPIE",

editor = "Branislav Vuksanovic and Jianhong Zhou and Antanas Verikas and Petia Radeva",

booktitle = "Seventh International Conference on Machine Vision, ICMV 2014",

address = "United States",

}

Sumer, O & Gunes, EO 2015, LBP and SIFT based facial expression recognition. in B Vuksanovic, J Zhou, A Verikas & P Radeva (eds), Seventh International Conference on Machine Vision, ICMV 2014., 94450A, Proceedings of SPIE - The International Society for Optical Engineering, vol. 9445, SPIE, 7th International Conference on Machine Vision, ICMV 2014, Milan, Italy, 19/11/14. https://doi.org/10.1117/12.2181505

LBP and SIFT based facial expression recognition. / Sumer, Omer; Gunes, Ece Olcay.
Seventh International Conference on Machine Vision, ICMV 2014. ed. / Branislav Vuksanovic; Jianhong Zhou; Antanas Verikas; Petia Radeva. SPIE, 2015. 94450A (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 9445).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - LBP and SIFT based facial expression recognition

AU - Sumer, Omer

AU - Gunes, Ece Olcay

PY - 2015

Y1 - 2015

N2 - This study compares the performance of local binary patterns (LBP) and scale invariant feature transform (SIFT) with support vector machines (SVM) in automatic classification of discrete facial expressions. Facial expression recognition is a multiclass classification problem and seven classes; happiness, anger, sadness, disgust, surprise, fear and comtempt are classified. Using SIFT feature vectors and linear SVM, 93.1% mean accuracy is acquired on CK+ database. On the other hand, the performance of LBP-based classifier with linear SVM is reported on SFEW using strictly person independent (SPI) protocol. Seven-class mean accuracy on SFEW is 59.76%. Experiments on both databases showed that LBP features can be used in a fairly descriptive way if a good localization of facial points and partitioning strategy are followed.

AB - This study compares the performance of local binary patterns (LBP) and scale invariant feature transform (SIFT) with support vector machines (SVM) in automatic classification of discrete facial expressions. Facial expression recognition is a multiclass classification problem and seven classes; happiness, anger, sadness, disgust, surprise, fear and comtempt are classified. Using SIFT feature vectors and linear SVM, 93.1% mean accuracy is acquired on CK+ database. On the other hand, the performance of LBP-based classifier with linear SVM is reported on SFEW using strictly person independent (SPI) protocol. Seven-class mean accuracy on SFEW is 59.76%. Experiments on both databases showed that LBP features can be used in a fairly descriptive way if a good localization of facial points and partitioning strategy are followed.

KW - emotion classification

KW - Facial Expression Recognition (FER)

KW - LBP

KW - SIFT

KW - SVM

UR - http://www.scopus.com/inward/record.url?scp=84924401997&partnerID=8YFLogxK

U2 - 10.1117/12.2181505

DO - 10.1117/12.2181505

M3 - Conference contribution

AN - SCOPUS:84924401997

T3 - Proceedings of SPIE - The International Society for Optical Engineering

BT - Seventh International Conference on Machine Vision, ICMV 2014

A2 - Vuksanovic, Branislav

A2 - Zhou, Jianhong

A2 - Verikas, Antanas

A2 - Radeva, Petia

PB - SPIE

T2 - 7th International Conference on Machine Vision, ICMV 2014

Y2 - 19 November 2014 through 21 November 2014

ER -

LBP and SIFT based facial expression recognition

Abstract

Publication series

Conference

Bibliographical note

Keywords

Access to Document

Other files and links

Fingerprint

Cite this