Enhancing 3D Facial Expression Recognition by, Exaggerating Geometry Characteristics

Weijian Li; Yunhong Wang; Huibin Li; Di Huang

doi:10.1007/978-3-319-69923-3_21

Enhancing 3D Facial Expression Recognition by, Exaggerating Geometry Characteristics

Weijian Li
, Yunhong Wang
, Huibin Li
, Di Huang

数学与统计学院

Beihang University

科研成果: 书/报告/会议事项章节 › 会议稿件 › 同行评审

摘要

This paper studies exaggerated facial shapes in addition to original facial shapes to assist 3D Facial Expression Recognition (FER). We propose a Poisson equation based approach to exaggerate facial shape characteristics to highlight expression clues that are latent in original facial surfaces but useful for recognizing expressions. To validate this idea, we exploit two off-the-shelf descriptors that reach state of the art performance in 3D FER, namely Geometric Scattering Representation (GSR) and Multi-Scale Local Normal Patterns (MS-LNPs) for expression-related feature extraction, and adopt early fusion to combine the credits of the original surface and the enhanced one, followed by the SVMs and Multiple Kernel Learning (MKL) classifiers. The accuracy gain of two features achieved on BU-3DFE is 0.8% and 1.3% respectively. Such results show that the exaggerated faces are complementary to the original faces in discriminating different facial expressions in the 3D domain.

源语言	英语
主期刊名	Biometric Recognition - 12th Chinese Conference, CCBR 2017, Proceedings
编辑	Yunhong Wang, Yu Qiao, Jie Zhou, Jianjiang Feng, Zhenan Sun, Zhenhua Guo, Shiguang Shan, Linlin Shen, Shiqi Yu, Yong Xu
出版商	Springer Verlag
页	191-200
页数	10
ISBN（印刷版）	9783319699226
DOI	https://doi.org/10.1007/978-3-319-69923-3_21
出版状态	已出版 - 2017
活动	12th Chinese Conference on Biometric Recognition, CCBR 2017 - Beijing, 中国期限: 28 10月 2017 → 29 10月 2017

出版系列

姓名	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
卷	10568 LNCS
ISSN（印刷版）	0302-9743
ISSN（电子版）	1611-3349

会议

会议	12th Chinese Conference on Biometric Recognition, CCBR 2017
国家/地区	中国
市	Beijing
时期	28/10/17 → 29/10/17

访问文件

10.1007/978-3-319-69923-3_21

其它文件与链接

链接到 Scopus 的出版物

引用此

Li, W., Wang, Y., Li, H., & Huang, D. (2017). Enhancing 3D Facial Expression Recognition by, Exaggerating Geometry Characteristics. 在 Y. Wang, Y. Qiao, J. Zhou, J. Feng, Z. Sun, Z. Guo, S. Shan, L. Shen, S. Yu, & Y. Xu (编辑), Biometric Recognition - 12th Chinese Conference, CCBR 2017, Proceedings (页码 191-200). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); 卷 10568 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-69923-3_21

Li, Weijian ; Wang, Yunhong ; Li, Huibin 等. / Enhancing 3D Facial Expression Recognition by, Exaggerating Geometry Characteristics. Biometric Recognition - 12th Chinese Conference, CCBR 2017, Proceedings. 编辑 / Yunhong Wang ; Yu Qiao ; Jie Zhou ; Jianjiang Feng ; Zhenan Sun ; Zhenhua Guo ; Shiguang Shan ; Linlin Shen ; Shiqi Yu ; Yong Xu. Springer Verlag, 2017. 页码 191-200 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{af8c7ebe43f345edbaa4bb569dbe4ffa,

title = "Enhancing 3D Facial Expression Recognition by, Exaggerating Geometry Characteristics",

abstract = "This paper studies exaggerated facial shapes in addition to original facial shapes to assist 3D Facial Expression Recognition (FER). We propose a Poisson equation based approach to exaggerate facial shape characteristics to highlight expression clues that are latent in original facial surfaces but useful for recognizing expressions. To validate this idea, we exploit two off-the-shelf descriptors that reach state of the art performance in 3D FER, namely Geometric Scattering Representation (GSR) and Multi-Scale Local Normal Patterns (MS-LNPs) for expression-related feature extraction, and adopt early fusion to combine the credits of the original surface and the enhanced one, followed by the SVMs and Multiple Kernel Learning (MKL) classifiers. The accuracy gain of two features achieved on BU-3DFE is 0.8\% and 1.3\% respectively. Such results show that the exaggerated faces are complementary to the original faces in discriminating different facial expressions in the 3D domain.",

keywords = "3D facial expression exaggeration, Facial Expression Recognition, Feature fusion",

author = "Weijian Li and Yunhong Wang and Huibin Li and Di Huang",

note = "Publisher Copyright: {\textcopyright} 2017, Springer International Publishing AG.; 12th Chinese Conference on Biometric Recognition, CCBR 2017 ; Conference date: 28-10-2017 Through 29-10-2017",

year = "2017",

doi = "10.1007/978-3-319-69923-3\_21",

language = "英语",

isbn = "9783319699226",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "191--200",

editor = "Yunhong Wang and Yu Qiao and Jie Zhou and Jianjiang Feng and Zhenan Sun and Zhenhua Guo and Shiguang Shan and Linlin Shen and Shiqi Yu and Yong Xu",

booktitle = "Biometric Recognition - 12th Chinese Conference, CCBR 2017, Proceedings",

}

Li, W, Wang, Y, Li, H & Huang, D 2017, Enhancing 3D Facial Expression Recognition by, Exaggerating Geometry Characteristics. 在 Y Wang, Y Qiao, J Zhou, J Feng, Z Sun, Z Guo, S Shan, L Shen, S Yu & Y Xu (编辑), Biometric Recognition - 12th Chinese Conference, CCBR 2017, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 卷 10568 LNCS, Springer Verlag, 页码 191-200, 12th Chinese Conference on Biometric Recognition, CCBR 2017, Beijing, 中国, 28/10/17. https://doi.org/10.1007/978-3-319-69923-3_21

Enhancing 3D Facial Expression Recognition by, Exaggerating Geometry Characteristics. / Li, Weijian; Wang, Yunhong; Li, Huibin 等.
Biometric Recognition - 12th Chinese Conference, CCBR 2017, Proceedings. 编辑 / Yunhong Wang; Yu Qiao; Jie Zhou; Jianjiang Feng; Zhenan Sun; Zhenhua Guo; Shiguang Shan; Linlin Shen; Shiqi Yu; Yong Xu. Springer Verlag, 2017. 页码 191-200 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); 卷 10568 LNCS).

科研成果: 书/报告/会议事项章节 › 会议稿件 › 同行评审

TY - GEN

T1 - Enhancing 3D Facial Expression Recognition by, Exaggerating Geometry Characteristics

AU - Li, Weijian

AU - Wang, Yunhong

AU - Li, Huibin

AU - Huang, Di

PY - 2017

Y1 - 2017

N2 - This paper studies exaggerated facial shapes in addition to original facial shapes to assist 3D Facial Expression Recognition (FER). We propose a Poisson equation based approach to exaggerate facial shape characteristics to highlight expression clues that are latent in original facial surfaces but useful for recognizing expressions. To validate this idea, we exploit two off-the-shelf descriptors that reach state of the art performance in 3D FER, namely Geometric Scattering Representation (GSR) and Multi-Scale Local Normal Patterns (MS-LNPs) for expression-related feature extraction, and adopt early fusion to combine the credits of the original surface and the enhanced one, followed by the SVMs and Multiple Kernel Learning (MKL) classifiers. The accuracy gain of two features achieved on BU-3DFE is 0.8% and 1.3% respectively. Such results show that the exaggerated faces are complementary to the original faces in discriminating different facial expressions in the 3D domain.

AB - This paper studies exaggerated facial shapes in addition to original facial shapes to assist 3D Facial Expression Recognition (FER). We propose a Poisson equation based approach to exaggerate facial shape characteristics to highlight expression clues that are latent in original facial surfaces but useful for recognizing expressions. To validate this idea, we exploit two off-the-shelf descriptors that reach state of the art performance in 3D FER, namely Geometric Scattering Representation (GSR) and Multi-Scale Local Normal Patterns (MS-LNPs) for expression-related feature extraction, and adopt early fusion to combine the credits of the original surface and the enhanced one, followed by the SVMs and Multiple Kernel Learning (MKL) classifiers. The accuracy gain of two features achieved on BU-3DFE is 0.8% and 1.3% respectively. Such results show that the exaggerated faces are complementary to the original faces in discriminating different facial expressions in the 3D domain.

KW - 3D facial expression exaggeration

KW - Facial Expression Recognition

KW - Feature fusion

UR - https://www.scopus.com/pages/publications/85032689475

U2 - 10.1007/978-3-319-69923-3_21

DO - 10.1007/978-3-319-69923-3_21

M3 - 会议稿件

AN - SCOPUS:85032689475

SN - 9783319699226

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 191

EP - 200

BT - Biometric Recognition - 12th Chinese Conference, CCBR 2017, Proceedings

A2 - Wang, Yunhong

A2 - Qiao, Yu

A2 - Zhou, Jie

A2 - Feng, Jianjiang

A2 - Sun, Zhenan

A2 - Guo, Zhenhua

A2 - Shan, Shiguang

A2 - Shen, Linlin

A2 - Yu, Shiqi

A2 - Xu, Yong

PB - Springer Verlag

T2 - 12th Chinese Conference on Biometric Recognition, CCBR 2017

Y2 - 28 October 2017 through 29 October 2017

ER -

Li W, Wang Y, Li H, Huang D. Enhancing 3D Facial Expression Recognition by, Exaggerating Geometry Characteristics. 在 Wang Y, Qiao Y, Zhou J, Feng J, Sun Z, Guo Z, Shan S, Shen L, Yu S, Xu Y, 编辑, Biometric Recognition - 12th Chinese Conference, CCBR 2017, Proceedings. Springer Verlag. 2017. 页码 191-200. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-69923-3_21

Enhancing 3D Facial Expression Recognition by, Exaggerating Geometry Characteristics

摘要

出版系列

会议

访问文件

其它文件与链接

学术指纹

引用此