Facial landmark detection under large pose

Yangyang Hao; Hengliang Zhu; Zhiwen Shao; Xin Tan; Lizhuang Ma

doi:10.1007/978-3-030-04212-7_60

Facial landmark detection under large pose

Yangyang Hao, Hengliang Zhu, Zhiwen Shao, Xin Tan, Lizhuang Ma

School of Computer Science and Technology

Shanghai Jiao Tong University

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

Abstract

Facial landmark detection is a necessary step in many vision tasks and plenty of excellent methods have been proposed to solve this problem. However, for the conditions with large pose and complex expression, these works usually suffer an eclipse. In this paper, we propose a two-stage cascade regression framework using patch-difference features to overcome the above problem. In the first stage, by applying the patch-difference feature and augmenting the large pose samples to the classical shape regression model, salient landmarks (eye centers, nose, mouth corners) can be located precisely. In the second stage, by applying enhanced feature section constraint to the patch-difference feature, multi-landmark detection is achieved. Experimental results show that our algorithm has a significant improvement compared to the classical shape regression method and achieves superior results on COFW dataset.

Original language	English
Title of host publication	Neural Information Processing - 25th International Conference, ICONIP 2018, Proceedings
Editors	Seiichi Ozawa, Andrew Chi Sing Leung, Long Cheng
Publisher	Springer Verlag
Pages	684-696
Number of pages	13
ISBN (Print)	9783030042110
DOIs	https://doi.org/10.1007/978-3-030-04212-7_60
State	Published - 2018
Event	25th International Conference on Neural Information Processing, ICONIP 2018 - Siem Reap, Cambodia Duration: 13 Dec 2018 → 16 Dec 2018

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11304 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	25th International Conference on Neural Information Processing, ICONIP 2018
Country/Territory	Cambodia
City	Siem Reap
Period	13/12/18 → 16/12/18

Keywords

Facial landmark detection
Feature section constraint
Large pose
Patch-difference feature

Access to Document

10.1007/978-3-030-04212-7_60

Cite this

Hao, Y., Zhu, H., Shao, Z., Tan, X., & Ma, L. (2018). Facial landmark detection under large pose. In S. Ozawa, A. C. S. Leung, & L. Cheng (Eds.), Neural Information Processing - 25th International Conference, ICONIP 2018, Proceedings (pp. 684-696). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11304 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-04212-7_60

Hao, Yangyang ; Zhu, Hengliang ; Shao, Zhiwen et al. / Facial landmark detection under large pose. Neural Information Processing - 25th International Conference, ICONIP 2018, Proceedings. editor / Seiichi Ozawa ; Andrew Chi Sing Leung ; Long Cheng. Springer Verlag, 2018. pp. 684-696 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{e2db54f84db3414a8f7b4ad60b972ea7,

title = "Facial landmark detection under large pose",

abstract = "Facial landmark detection is a necessary step in many vision tasks and plenty of excellent methods have been proposed to solve this problem. However, for the conditions with large pose and complex expression, these works usually suffer an eclipse. In this paper, we propose a two-stage cascade regression framework using patch-difference features to overcome the above problem. In the first stage, by applying the patch-difference feature and augmenting the large pose samples to the classical shape regression model, salient landmarks (eye centers, nose, mouth corners) can be located precisely. In the second stage, by applying enhanced feature section constraint to the patch-difference feature, multi-landmark detection is achieved. Experimental results show that our algorithm has a significant improvement compared to the classical shape regression method and achieves superior results on COFW dataset.",

keywords = "Facial landmark detection, Feature section constraint, Large pose, Patch-difference feature",

author = "Yangyang Hao and Hengliang Zhu and Zhiwen Shao and Xin Tan and Lizhuang Ma",

note = "Publisher Copyright: {\textcopyright} Springer Nature Switzerland AG 2018.; 25th International Conference on Neural Information Processing, ICONIP 2018 ; Conference date: 13-12-2018 Through 16-12-2018",

year = "2018",

doi = "10.1007/978-3-030-04212-7\_60",

language = "英语",

isbn = "9783030042110",

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

publisher = "Springer Verlag",

pages = "684--696",

editor = "Seiichi Ozawa and Leung, \{Andrew Chi Sing\} and Long Cheng",

booktitle = "Neural Information Processing - 25th International Conference, ICONIP 2018, Proceedings",

address = "德国",

}

Hao, Y, Zhu, H, Shao, Z, Tan, X & Ma, L 2018, Facial landmark detection under large pose. in S Ozawa, ACS Leung & L Cheng (eds), Neural Information Processing - 25th International Conference, ICONIP 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11304 LNCS, Springer Verlag, pp. 684-696, 25th International Conference on Neural Information Processing, ICONIP 2018, Siem Reap, Cambodia, 13/12/18. https://doi.org/10.1007/978-3-030-04212-7_60

Facial landmark detection under large pose. / Hao, Yangyang; Zhu, Hengliang; Shao, Zhiwen et al.
Neural Information Processing - 25th International Conference, ICONIP 2018, Proceedings. ed. / Seiichi Ozawa; Andrew Chi Sing Leung; Long Cheng. Springer Verlag, 2018. p. 684-696 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11304 LNCS).

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

TY - GEN

T1 - Facial landmark detection under large pose

AU - Hao, Yangyang

AU - Zhu, Hengliang

AU - Shao, Zhiwen

AU - Tan, Xin

AU - Ma, Lizhuang

N1 - Publisher Copyright: © Springer Nature Switzerland AG 2018.

PY - 2018

Y1 - 2018

N2 - Facial landmark detection is a necessary step in many vision tasks and plenty of excellent methods have been proposed to solve this problem. However, for the conditions with large pose and complex expression, these works usually suffer an eclipse. In this paper, we propose a two-stage cascade regression framework using patch-difference features to overcome the above problem. In the first stage, by applying the patch-difference feature and augmenting the large pose samples to the classical shape regression model, salient landmarks (eye centers, nose, mouth corners) can be located precisely. In the second stage, by applying enhanced feature section constraint to the patch-difference feature, multi-landmark detection is achieved. Experimental results show that our algorithm has a significant improvement compared to the classical shape regression method and achieves superior results on COFW dataset.

AB - Facial landmark detection is a necessary step in many vision tasks and plenty of excellent methods have been proposed to solve this problem. However, for the conditions with large pose and complex expression, these works usually suffer an eclipse. In this paper, we propose a two-stage cascade regression framework using patch-difference features to overcome the above problem. In the first stage, by applying the patch-difference feature and augmenting the large pose samples to the classical shape regression model, salient landmarks (eye centers, nose, mouth corners) can be located precisely. In the second stage, by applying enhanced feature section constraint to the patch-difference feature, multi-landmark detection is achieved. Experimental results show that our algorithm has a significant improvement compared to the classical shape regression method and achieves superior results on COFW dataset.

KW - Facial landmark detection

KW - Feature section constraint

KW - Large pose

KW - Patch-difference feature

UR - https://www.scopus.com/pages/publications/85058983409

U2 - 10.1007/978-3-030-04212-7_60

DO - 10.1007/978-3-030-04212-7_60

M3 - 会议稿件

AN - SCOPUS:85058983409

SN - 9783030042110

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

SP - 684

EP - 696

BT - Neural Information Processing - 25th International Conference, ICONIP 2018, Proceedings

A2 - Ozawa, Seiichi

A2 - Leung, Andrew Chi Sing

A2 - Cheng, Long

PB - Springer Verlag

T2 - 25th International Conference on Neural Information Processing, ICONIP 2018

Y2 - 13 December 2018 through 16 December 2018

ER -

Hao Y, Zhu H, Shao Z, Tan X, Ma L. Facial landmark detection under large pose. In Ozawa S, Leung ACS, Cheng L, editors, Neural Information Processing - 25th International Conference, ICONIP 2018, Proceedings. Springer Verlag. 2018. p. 684-696. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-04212-7_60

Facial landmark detection under large pose

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this