Deep cascaded Bi-network for face hallucination

Shizhan Zhu; Sifei Liu; Chen Change Loy; Xiaoou Tang

doi:10.1007/978-3-319-46454-1_37

Deep cascaded Bi-network for face hallucination

Shizhan Zhu, Sifei Liu, Chen Change Loy^*, Xiaoou Tang

^*Corresponding author for this work

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

189 Citations (Scopus)

Abstract

We present a novel framework for hallucinating faces of unconstrained poses and with very low resolution (face size as small as 5pxIOD). In contrast to existing studies that mostly ignore or assume pre-aligned face spatial configuration (e.g. facial landmarks localization or dense correspondence field), we alternatingly optimize two complementary tasks, namely face hallucination and dense correspondence field estimation, in a unified framework. In addition, we propose a new gated deep bi-network that contains two functionality-specialized branches to recover different levels of texture details. Extensive experiments demonstrate that such formulation allows exceptional hallucination quality on in-the-wild low-res faces with significant pose and illumination variations.

Original language	English
Title of host publication	Computer Vision - 14th European Conference, ECCV 2016, Proceedings
Editors	Bastian Leibe, Jiri Matas, Nicu Sebe, Max Welling
Publisher	Springer Verlag
Pages	614-630
Number of pages	17
ISBN (Print)	9783319464534
DOIs	https://doi.org/10.1007/978-3-319-46454-1_37
Publication status	Published - 2016
Externally published	Yes
Event	14th European Conference on Computer Vision, ECCV 2016 - Amsterdam, Netherlands Duration: Oct 8 2016 → Oct 16 2016

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	9909 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	14th European Conference on Computer Vision, ECCV 2016
Country/Territory	Netherlands
City	Amsterdam
Period	10/8/16 → 10/16/16

Bibliographical note

Publisher Copyright:
© Springer International Publishing AG 2016.

ASJC Scopus Subject Areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-319-46454-1_37

Cite this

Zhu, S., Liu, S., Loy, C. C., & Tang, X. (2016). Deep cascaded Bi-network for face hallucination. In B. Leibe, J. Matas, N. Sebe, & M. Welling (Eds.), Computer Vision - 14th European Conference, ECCV 2016, Proceedings (pp. 614-630). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9909 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-46454-1_37

Zhu, Shizhan ; Liu, Sifei ; Loy, Chen Change et al. / Deep cascaded Bi-network for face hallucination. Computer Vision - 14th European Conference, ECCV 2016, Proceedings. editor / Bastian Leibe ; Jiri Matas ; Nicu Sebe ; Max Welling. Springer Verlag, 2016. pp. 614-630 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{5568e70825f147e89eaaa76204040a2d,

title = "Deep cascaded Bi-network for face hallucination",

abstract = "We present a novel framework for hallucinating faces of unconstrained poses and with very low resolution (face size as small as 5pxIOD). In contrast to existing studies that mostly ignore or assume pre-aligned face spatial configuration (e.g. facial landmarks localization or dense correspondence field), we alternatingly optimize two complementary tasks, namely face hallucination and dense correspondence field estimation, in a unified framework. In addition, we propose a new gated deep bi-network that contains two functionality-specialized branches to recover different levels of texture details. Extensive experiments demonstrate that such formulation allows exceptional hallucination quality on in-the-wild low-res faces with significant pose and illumination variations.",

author = "Shizhan Zhu and Sifei Liu and Loy, \{Chen Change\} and Xiaoou Tang",

note = "Publisher Copyright: {\textcopyright} Springer International Publishing AG 2016.; 14th European Conference on Computer Vision, ECCV 2016 ; Conference date: 08-10-2016 Through 16-10-2016",

year = "2016",

doi = "10.1007/978-3-319-46454-1\_37",

language = "English",

isbn = "9783319464534",

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

publisher = "Springer Verlag",

pages = "614--630",

editor = "Bastian Leibe and Jiri Matas and Nicu Sebe and Max Welling",

booktitle = "Computer Vision - 14th European Conference, ECCV 2016, Proceedings",

address = "Germany",

}

Zhu, S, Liu, S, Loy, CC & Tang, X 2016, Deep cascaded Bi-network for face hallucination. in B Leibe, J Matas, N Sebe & M Welling (eds), Computer Vision - 14th European Conference, ECCV 2016, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9909 LNCS, Springer Verlag, pp. 614-630, 14th European Conference on Computer Vision, ECCV 2016, Amsterdam, Netherlands, 10/8/16. https://doi.org/10.1007/978-3-319-46454-1_37

Deep cascaded Bi-network for face hallucination. / Zhu, Shizhan; Liu, Sifei; Loy, Chen Change et al.
Computer Vision - 14th European Conference, ECCV 2016, Proceedings. ed. / Bastian Leibe; Jiri Matas; Nicu Sebe; Max Welling. Springer Verlag, 2016. p. 614-630 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9909 LNCS).

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

TY - GEN

T1 - Deep cascaded Bi-network for face hallucination

AU - Zhu, Shizhan

AU - Liu, Sifei

AU - Loy, Chen Change

AU - Tang, Xiaoou

N1 - Publisher Copyright: © Springer International Publishing AG 2016.

PY - 2016

Y1 - 2016

N2 - We present a novel framework for hallucinating faces of unconstrained poses and with very low resolution (face size as small as 5pxIOD). In contrast to existing studies that mostly ignore or assume pre-aligned face spatial configuration (e.g. facial landmarks localization or dense correspondence field), we alternatingly optimize two complementary tasks, namely face hallucination and dense correspondence field estimation, in a unified framework. In addition, we propose a new gated deep bi-network that contains two functionality-specialized branches to recover different levels of texture details. Extensive experiments demonstrate that such formulation allows exceptional hallucination quality on in-the-wild low-res faces with significant pose and illumination variations.

AB - We present a novel framework for hallucinating faces of unconstrained poses and with very low resolution (face size as small as 5pxIOD). In contrast to existing studies that mostly ignore or assume pre-aligned face spatial configuration (e.g. facial landmarks localization or dense correspondence field), we alternatingly optimize two complementary tasks, namely face hallucination and dense correspondence field estimation, in a unified framework. In addition, we propose a new gated deep bi-network that contains two functionality-specialized branches to recover different levels of texture details. Extensive experiments demonstrate that such formulation allows exceptional hallucination quality on in-the-wild low-res faces with significant pose and illumination variations.

UR - http://www.scopus.com/inward/record.url?scp=84990050338&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84990050338&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-46454-1_37

DO - 10.1007/978-3-319-46454-1_37

M3 - Conference contribution

AN - SCOPUS:84990050338

SN - 9783319464534

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

SP - 614

EP - 630

BT - Computer Vision - 14th European Conference, ECCV 2016, Proceedings

A2 - Leibe, Bastian

A2 - Matas, Jiri

A2 - Sebe, Nicu

A2 - Welling, Max

PB - Springer Verlag

T2 - 14th European Conference on Computer Vision, ECCV 2016

Y2 - 8 October 2016 through 16 October 2016

ER -

Zhu S, Liu S, Loy CC, Tang X. Deep cascaded Bi-network for face hallucination. In Leibe B, Matas J, Sebe N, Welling M, editors, Computer Vision - 14th European Conference, ECCV 2016, Proceedings. Springer Verlag. 2016. p. 614-630. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-319-46454-1_37

Deep cascaded Bi-network for face hallucination

Abstract

Publication series

Conference

Bibliographical note

ASJC Scopus Subject Areas

Access to Document

Other files and links

Cite this