Facial analysis with a Lie group kernel

Chunyan Xu, Canyi Lu, Junbin Gao, Tianjiang Wang, Shuicheng Yan

Research output: Contribution to journalArticle

4 Citations (Scopus)
4 Downloads (Pure)

Abstract

To efficiently deal with the complex nonlinear variations of face images, a novel Lie group kernel is proposed in this work to address the facial analysis problems. Firstly, we present a linear dynamic model (LDM) based face representation to capture both the appearance and spatial information of the face image. Secondly, the derived linear dynamic model can be parameterized as a specially-structured upper triangular matrix, the space of which is proved to constitute a Lie group. A Lie group (LG) kernel is then designed to characterize the similarity between the linear dynamic models for any two face images and the kernel can be fed into classical kernel-based classifiers for different types of facial analysis. Finally, experimental evaluations on face recognition and head pose estimation are conducted on several challenging datasets and the results show that the proposed algorithm outperforms other facial analysis methods.
Original languageEnglish
Pages (from-to)1-11
Number of pages11
JournalIEEE Transactions on Circuits and Systems for Video Technology
Volume25
Issue number7
Early online dateOct 2014
DOIs
Publication statusPublished - Jul 2015

Fingerprint

Lie groups
Dynamic models
Face recognition
Classifiers

Cite this

Xu, Chunyan ; Lu, Canyi ; Gao, Junbin ; Wang, Tianjiang ; Yan, Shuicheng. / Facial analysis with a Lie group kernel. In: IEEE Transactions on Circuits and Systems for Video Technology. 2015 ; Vol. 25, No. 7. pp. 1-11.
@article{0a679c69b34d42d584215479650d813b,
title = "Facial analysis with a Lie group kernel",
abstract = "To efficiently deal with the complex nonlinear variations of face images, a novel Lie group kernel is proposed in this work to address the facial analysis problems. Firstly, we present a linear dynamic model (LDM) based face representation to capture both the appearance and spatial information of the face image. Secondly, the derived linear dynamic model can be parameterized as a specially-structured upper triangular matrix, the space of which is proved to constitute a Lie group. A Lie group (LG) kernel is then designed to characterize the similarity between the linear dynamic models for any two face images and the kernel can be fed into classical kernel-based classifiers for different types of facial analysis. Finally, experimental evaluations on face recognition and head pose estimation are conducted on several challenging datasets and the results show that the proposed algorithm outperforms other facial analysis methods.",
keywords = "Facial analysis, Kernel learning, Lie group (LG) manifold",
author = "Chunyan Xu and Canyi Lu and Junbin Gao and Tianjiang Wang and Shuicheng Yan",
note = "Includes bibliographical references.",
year = "2015",
month = "7",
doi = "10.1109/TCSVT.2014.2365655",
language = "English",
volume = "25",
pages = "1--11",
journal = "IEEE Transactions on Circuits and Systems for Video Technology",
issn = "1051-8215",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",
number = "7",

}

Facial analysis with a Lie group kernel. / Xu, Chunyan; Lu, Canyi; Gao, Junbin; Wang, Tianjiang; Yan, Shuicheng.

In: IEEE Transactions on Circuits and Systems for Video Technology, Vol. 25, No. 7, 07.2015, p. 1-11.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Facial analysis with a Lie group kernel

AU - Xu, Chunyan

AU - Lu, Canyi

AU - Gao, Junbin

AU - Wang, Tianjiang

AU - Yan, Shuicheng

N1 - Includes bibliographical references.

PY - 2015/7

Y1 - 2015/7

N2 - To efficiently deal with the complex nonlinear variations of face images, a novel Lie group kernel is proposed in this work to address the facial analysis problems. Firstly, we present a linear dynamic model (LDM) based face representation to capture both the appearance and spatial information of the face image. Secondly, the derived linear dynamic model can be parameterized as a specially-structured upper triangular matrix, the space of which is proved to constitute a Lie group. A Lie group (LG) kernel is then designed to characterize the similarity between the linear dynamic models for any two face images and the kernel can be fed into classical kernel-based classifiers for different types of facial analysis. Finally, experimental evaluations on face recognition and head pose estimation are conducted on several challenging datasets and the results show that the proposed algorithm outperforms other facial analysis methods.

AB - To efficiently deal with the complex nonlinear variations of face images, a novel Lie group kernel is proposed in this work to address the facial analysis problems. Firstly, we present a linear dynamic model (LDM) based face representation to capture both the appearance and spatial information of the face image. Secondly, the derived linear dynamic model can be parameterized as a specially-structured upper triangular matrix, the space of which is proved to constitute a Lie group. A Lie group (LG) kernel is then designed to characterize the similarity between the linear dynamic models for any two face images and the kernel can be fed into classical kernel-based classifiers for different types of facial analysis. Finally, experimental evaluations on face recognition and head pose estimation are conducted on several challenging datasets and the results show that the proposed algorithm outperforms other facial analysis methods.

KW - Facial analysis

KW - Kernel learning

KW - Lie group (LG) manifold

U2 - 10.1109/TCSVT.2014.2365655

DO - 10.1109/TCSVT.2014.2365655

M3 - Article

VL - 25

SP - 1

EP - 11

JO - IEEE Transactions on Circuits and Systems for Video Technology

JF - IEEE Transactions on Circuits and Systems for Video Technology

SN - 1051-8215

IS - 7

ER -