Detecting people in artwork with CNNs

Nicholas Westlake; Hongping Cai; Peter Hall

doi:10.1007/978-3-319-46604-0_57

Detecting people in artwork with CNNs

Nicholas Westlake, Hongping Cai, Peter Hall

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

11 Citations (Scopus)

Abstract

CNNs have massively improved performance in object detection in photographs. However research into object detection in artwork remains limited. We show state-of-the-art performance on a challenging dataset, People-Art, which contains people from photos, cartoons and 41 different artwork movements. We achieve this high performance by finetuning a CNN for this task, thus also demonstrating that training CNNs on photos results in overfitting for photos: only the first three or four layers transfer from photos to artwork. Although the CNN’s performance is the highest yet, it remains less than 60% AP, suggesting further work is needed for the cross-depiction problem.

Original language	English
Title of host publication	Computer Vision - ECCV 2016 Workshops. ECCV 2016.
Editors	Gang Hua, Hervé Jégou
Publisher	Springer Verlag
Pages	825-841
Number of pages	17
ISBN (Print)	9783319466033
DOIs	https://doi.org/10.1007/978-3-319-46604-0_57
Publication status	Published - 18 Sep 2016
Event	14th European Conference on Computer Vision, ECCV 2016 - Amsterdam, Netherlands Duration: 8 Oct 2016 → 16 Oct 2016

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	9913 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

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

Keywords

CNNs
Cross-depiction problem
Object recognition

ASJC Scopus subject areas

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/978-3-319-46604-0_57

Link to publication in Scopus

Projects

1 Active

Centre for the Analysis of Motion, Entertainment Research and Applications (CAMERA)

Cosker, D., Campbell, N., Fincham Haines, T., Hall, P., Kim, K. I., Lutteroth, C., O'Neill, E., Richardt, C. & Yang, Y.

Engineering and Physical Sciences Research Council

1/09/15 → 28/02/21

Project: Research council

Cite this

Westlake, N., Cai, H., & Hall, P. (2016). Detecting people in artwork with CNNs. In G. Hua, & H. Jégou (Eds.), Computer Vision - ECCV 2016 Workshops. ECCV 2016. (pp. 825-841). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9913 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-46604-0_57

Detecting people in artwork with CNNs. / Westlake, Nicholas; Cai, Hongping; Hall, Peter.

Computer Vision - ECCV 2016 Workshops. ECCV 2016.. ed. / Gang Hua; Hervé Jégou. Springer Verlag, 2016. p. 825-841 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 9913 LNCS).

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

Westlake, N, Cai, H & Hall, P 2016, Detecting people in artwork with CNNs. in G Hua & H Jégou (eds), Computer Vision - ECCV 2016 Workshops. ECCV 2016.. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 9913 LNCS, Springer Verlag, pp. 825-841, 14th European Conference on Computer Vision, ECCV 2016, Amsterdam, Netherlands, 8/10/16. https://doi.org/10.1007/978-3-319-46604-0_57

@inproceedings{c50337d1210a48ca8ad9bd3ce29e6ce4,

title = "Detecting people in artwork with CNNs",

abstract = "CNNs have massively improved performance in object detection in photographs. However research into object detection in artwork remains limited. We show state-of-the-art performance on a challenging dataset, People-Art, which contains people from photos, cartoons and 41 different artwork movements. We achieve this high performance by finetuning a CNN for this task, thus also demonstrating that training CNNs on photos results in overfitting for photos: only the first three or four layers transfer from photos to artwork. Although the CNN{\textquoteright}s performance is the highest yet, it remains less than 60% AP, suggesting further work is needed for the cross-depiction problem.",

keywords = "CNNs, Cross-depiction problem, Object recognition",

author = "Nicholas Westlake and Hongping Cai and Peter Hall",

year = "2016",

month = sep,

day = "18",

doi = "10.1007/978-3-319-46604-0_57",

language = "English",

isbn = "9783319466033",

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

publisher = "Springer Verlag",

pages = "825--841",

editor = "Gang Hua and Herv{\'e} J{\'e}gou",

booktitle = "Computer Vision - ECCV 2016 Workshops. ECCV 2016.",

note = "14th European Conference on Computer Vision, ECCV 2016 ; Conference date: 08-10-2016 Through 16-10-2016",

}

TY - GEN

T1 - Detecting people in artwork with CNNs

AU - Westlake, Nicholas

AU - Cai, Hongping

AU - Hall, Peter

PY - 2016/9/18

Y1 - 2016/9/18

N2 - CNNs have massively improved performance in object detection in photographs. However research into object detection in artwork remains limited. We show state-of-the-art performance on a challenging dataset, People-Art, which contains people from photos, cartoons and 41 different artwork movements. We achieve this high performance by finetuning a CNN for this task, thus also demonstrating that training CNNs on photos results in overfitting for photos: only the first three or four layers transfer from photos to artwork. Although the CNN’s performance is the highest yet, it remains less than 60% AP, suggesting further work is needed for the cross-depiction problem.

AB - CNNs have massively improved performance in object detection in photographs. However research into object detection in artwork remains limited. We show state-of-the-art performance on a challenging dataset, People-Art, which contains people from photos, cartoons and 41 different artwork movements. We achieve this high performance by finetuning a CNN for this task, thus also demonstrating that training CNNs on photos results in overfitting for photos: only the first three or four layers transfer from photos to artwork. Although the CNN’s performance is the highest yet, it remains less than 60% AP, suggesting further work is needed for the cross-depiction problem.

KW - CNNs

KW - Cross-depiction problem

KW - Object recognition

UR - http://www.scopus.com/inward/record.url?scp=84989834205&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-46604-0_57

DO - 10.1007/978-3-319-46604-0_57

M3 - Conference contribution

AN - SCOPUS:84989834205

SN - 9783319466033

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

SP - 825

EP - 841

BT - Computer Vision - ECCV 2016 Workshops. ECCV 2016.

A2 - Hua, Gang

A2 - Jégou, Hervé

PB - Springer Verlag

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

Y2 - 8 October 2016 through 16 October 2016

ER -