Continuous supervised descent method for facial landmark localisation

Marc Oliu; Ciprian Corneanu; László A. Jeni; Jeffrey F. Cohn; Takeo Kanade; Sergio Escalera

doi:10.1007/978-3-319-54184-6_8

Continuous supervised descent method for facial landmark localisation

Marc Oliu^*, Ciprian Corneanu, László A. Jeni, Jeffrey F. Cohn, Takeo Kanade, Sergio Escalera

^*Corresponding author for this work

Research output: Contribution to book/anthology/report/conference proceeding › Article in proceeding › Research › peer-review

1 Citation (Scopus)

Abstract

Recent methods for facial landmark location perform well on close-to-frontal faces but have problems in generalising to large head rotations. In order to address this issue we propose a second order linear regression method that is both compact and robust against strong rotations. We provide a closed form solution, making the method fast to train. We test the method’s performance on two challenging datasets. The first has been intensely used by the community. The second has been specially generated from a well known 3D face dataset. It is considerably more challenging, including a high diversity of rotations and more samples than any other existing public dataset. The proposed method is compared against state-of-the-art approaches, including RCPR, CGPRT, LBF, CFSS, and GSDM. Results upon both datasets show that the proposed method offers state-of-the-art performance on near frontal view data, improves state-of-the-art methods on more challenging head rotation problems and keeps a compact model size.

Original language	English
Title of host publication	Computer Vision - ACCV 2016 - 13th Asian Conference on Computer Vision, Revised Selected Papers
Editors	Yoichi Sato, Shang-Hong Lai, Vincent Lepetit, Ko Nishino
Number of pages	15
Publisher	Physica-Verlag
Publication date	2017
Pages	121-135
ISBN (Print)	9783319541839
DOIs	https://doi.org/10.1007/978-3-319-54184-6_8
Publication status	Published - 2017
Event	13th Asian Conference on Computer Vision, ACCV 2016 - Taipei, Taiwan, Province of China Duration: 20 Nov 2016 → 24 Nov 2016

Conference

Conference	13th Asian Conference on Computer Vision, ACCV 2016
Country/Territory	Taiwan, Province of China
City	Taipei
Period	20/11/2016 → 24/11/2016

Series	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	10112 LNCS
ISSN	0302-9743

Bibliographical note

Funding Information:
The work of Marc Oliu is supported by the FI-DGR 2016 fellowship, granted by the Universities and Research Secretary of the Knowledge and Economy Department of the Generalitat de Catalunya. This work has been partially supported by the Spanish project TIN2013-43478-P, the European Comission Horizon 2020 granted project SEE.4C under call H2020-ICT-2015 and the U.S. National Institutes of Health under the grant MH096951.

Publisher Copyright:
© Springer International Publishing AG 2017.

Access to Document

10.1007/978-3-319-54184-6_8

AUB Link

Search for the material in Aalborg University Library's search engine

Cite this

Oliu, M., Corneanu, C., Jeni, L. A., Cohn, J. F., Kanade, T., & Escalera, S. (2017). Continuous supervised descent method for facial landmark localisation. In Y. Sato, S.-H. Lai, V. Lepetit, & K. Nishino (Eds.), Computer Vision - ACCV 2016 - 13th Asian Conference on Computer Vision, Revised Selected Papers (pp. 121-135). Physica-Verlag. https://doi.org/10.1007/978-3-319-54184-6_8

Oliu, Marc ; Corneanu, Ciprian ; Jeni, László A. et al. / Continuous supervised descent method for facial landmark localisation. Computer Vision - ACCV 2016 - 13th Asian Conference on Computer Vision, Revised Selected Papers. editor / Yoichi Sato ; Shang-Hong Lai ; Vincent Lepetit ; Ko Nishino. Physica-Verlag, 2017. pp. 121-135 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 10112 LNCS).

@inproceedings{f396eea949514beeaec0cc84c43d2aeb,

title = "Continuous supervised descent method for facial landmark localisation",

abstract = "Recent methods for facial landmark location perform well on close-to-frontal faces but have problems in generalising to large head rotations. In order to address this issue we propose a second order linear regression method that is both compact and robust against strong rotations. We provide a closed form solution, making the method fast to train. We test the method{\textquoteright}s performance on two challenging datasets. The first has been intensely used by the community. The second has been specially generated from a well known 3D face dataset. It is considerably more challenging, including a high diversity of rotations and more samples than any other existing public dataset. The proposed method is compared against state-of-the-art approaches, including RCPR, CGPRT, LBF, CFSS, and GSDM. Results upon both datasets show that the proposed method offers state-of-the-art performance on near frontal view data, improves state-of-the-art methods on more challenging head rotation problems and keeps a compact model size.",

author = "Marc Oliu and Ciprian Corneanu and Jeni, {L{\'a}szl{\'o} A.} and Cohn, {Jeffrey F.} and Takeo Kanade and Sergio Escalera",

note = "Funding Information: The work of Marc Oliu is supported by the FI-DGR 2016 fellowship, granted by the Universities and Research Secretary of the Knowledge and Economy Department of the Generalitat de Catalunya. This work has been partially supported by the Spanish project TIN2013-43478-P, the European Comission Horizon 2020 granted project SEE.4C under call H2020-ICT-2015 and the U.S. National Institutes of Health under the grant MH096951. Publisher Copyright: {\textcopyright} Springer International Publishing AG 2017.; 13th Asian Conference on Computer Vision, ACCV 2016 ; Conference date: 20-11-2016 Through 24-11-2016",

year = "2017",

doi = "10.1007/978-3-319-54184-6_8",

language = "English",

isbn = "9783319541839",

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

publisher = "Physica-Verlag",

pages = "121--135",

editor = "Yoichi Sato and Shang-Hong Lai and Vincent Lepetit and Ko Nishino",

booktitle = "Computer Vision - ACCV 2016 - 13th Asian Conference on Computer Vision, Revised Selected Papers",

}

Oliu, M, Corneanu, C, Jeni, LA, Cohn, JF, Kanade, T & Escalera, S 2017, Continuous supervised descent method for facial landmark localisation. in Y Sato, S-H Lai, V Lepetit & K Nishino (eds), Computer Vision - ACCV 2016 - 13th Asian Conference on Computer Vision, Revised Selected Papers. Physica-Verlag, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10112 LNCS, pp. 121-135, 13th Asian Conference on Computer Vision, ACCV 2016, Taipei, Taiwan, Province of China, 20/11/2016. https://doi.org/10.1007/978-3-319-54184-6_8

Continuous supervised descent method for facial landmark localisation. / Oliu, Marc; Corneanu, Ciprian; Jeni, László A. et al.
Computer Vision - ACCV 2016 - 13th Asian Conference on Computer Vision, Revised Selected Papers. ed. / Yoichi Sato; Shang-Hong Lai; Vincent Lepetit; Ko Nishino. Physica-Verlag, 2017. p. 121-135 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 10112 LNCS).

Research output: Contribution to book/anthology/report/conference proceeding › Article in proceeding › Research › peer-review

TY - GEN

T1 - Continuous supervised descent method for facial landmark localisation

AU - Oliu, Marc

AU - Corneanu, Ciprian

AU - Jeni, László A.

AU - Cohn, Jeffrey F.

AU - Kanade, Takeo

AU - Escalera, Sergio

N1 - Funding Information: The work of Marc Oliu is supported by the FI-DGR 2016 fellowship, granted by the Universities and Research Secretary of the Knowledge and Economy Department of the Generalitat de Catalunya. This work has been partially supported by the Spanish project TIN2013-43478-P, the European Comission Horizon 2020 granted project SEE.4C under call H2020-ICT-2015 and the U.S. National Institutes of Health under the grant MH096951. Publisher Copyright: © Springer International Publishing AG 2017.

PY - 2017

Y1 - 2017

N2 - Recent methods for facial landmark location perform well on close-to-frontal faces but have problems in generalising to large head rotations. In order to address this issue we propose a second order linear regression method that is both compact and robust against strong rotations. We provide a closed form solution, making the method fast to train. We test the method’s performance on two challenging datasets. The first has been intensely used by the community. The second has been specially generated from a well known 3D face dataset. It is considerably more challenging, including a high diversity of rotations and more samples than any other existing public dataset. The proposed method is compared against state-of-the-art approaches, including RCPR, CGPRT, LBF, CFSS, and GSDM. Results upon both datasets show that the proposed method offers state-of-the-art performance on near frontal view data, improves state-of-the-art methods on more challenging head rotation problems and keeps a compact model size.

AB - Recent methods for facial landmark location perform well on close-to-frontal faces but have problems in generalising to large head rotations. In order to address this issue we propose a second order linear regression method that is both compact and robust against strong rotations. We provide a closed form solution, making the method fast to train. We test the method’s performance on two challenging datasets. The first has been intensely used by the community. The second has been specially generated from a well known 3D face dataset. It is considerably more challenging, including a high diversity of rotations and more samples than any other existing public dataset. The proposed method is compared against state-of-the-art approaches, including RCPR, CGPRT, LBF, CFSS, and GSDM. Results upon both datasets show that the proposed method offers state-of-the-art performance on near frontal view data, improves state-of-the-art methods on more challenging head rotation problems and keeps a compact model size.

UR - http://www.scopus.com/inward/record.url?scp=85016206409&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-54184-6_8

DO - 10.1007/978-3-319-54184-6_8

M3 - Article in proceeding

AN - SCOPUS:85016206409

SN - 9783319541839

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

SP - 121

EP - 135

BT - Computer Vision - ACCV 2016 - 13th Asian Conference on Computer Vision, Revised Selected Papers

A2 - Sato, Yoichi

A2 - Lai, Shang-Hong

A2 - Lepetit, Vincent

A2 - Nishino, Ko

PB - Physica-Verlag

T2 - 13th Asian Conference on Computer Vision, ACCV 2016

Y2 - 20 November 2016 through 24 November 2016

ER -

Oliu M, Corneanu C, Jeni LA, Cohn JF, Kanade T, Escalera S. Continuous supervised descent method for facial landmark localisation. In Sato Y, Lai SH, Lepetit V, Nishino K, editors, Computer Vision - ACCV 2016 - 13th Asian Conference on Computer Vision, Revised Selected Papers. Physica-Verlag. 2017. p. 121-135. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), Vol. 10112 LNCS). doi: 10.1007/978-3-319-54184-6_8

Continuous supervised descent method for facial landmark localisation

Abstract

Conference

Bibliographical note

Access to Document

AUB Link

Other files and links

Fingerprint

Cite this