Large-scale object classification using label relation graphs

Jia Deng; Nan Ding; Yangqing Jia; Andrea Frome; Kevin Murphy; Samy Bengio; Yuan Li; Hartmut Neven; Hartwig Adam

doi:10.1007/978-3-319-10590-1_4

Large-scale object classification using label relation graphs

Jia Deng, Nan Ding, Yangqing Jia, Andrea Frome, Kevin Murphy, Samy Bengio, Yuan Li, Hartmut Neven, Hartwig Adam

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

267 Scopus citations

Abstract

In this paper we study how to perform object classification in a principled way that exploits the rich structure of real world labels. We develop a new model that allows encoding of flexible relations between labels. We introduce Hierarchy and Exclusion (HEX) graphs, a new formalism that captures semantic relations between any two labels applied to the same object: mutual exclusion, overlap and subsumption. We then provide rigorous theoretical analysis that illustrates properties of HEX graphs such as consistency, equivalence, and computational implications of the graph structure. Next, we propose a probabilistic classification model based on HEX graphs and show that it enjoys a number of desirable properties. Finally, we evaluate our method using a large-scale benchmark. Empirical results demonstrate that our model can significantly improve object classification by exploiting the label relations.

Original language	English (US)
Title of host publication	Computer Vision, ECCV 2014 - 13th European Conference, Proceedings
Publisher	Springer Verlag
Pages	48-64
Number of pages	17
Edition	PART 1
ISBN (Print)	9783319105895
DOIs	https://doi.org/10.1007/978-3-319-10590-1_4
State	Published - 2014
Externally published	Yes
Event	13th European Conference on Computer Vision, ECCV 2014 - Zurich, Switzerland Duration: Sep 6 2014 → Sep 12 2014

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 1
Volume	8689 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Other

Other	13th European Conference on Computer Vision, ECCV 2014
Country/Territory	Switzerland
City	Zurich
Period	9/6/14 → 9/12/14

All Science Journal Classification (ASJC) codes

Theoretical Computer Science
General Computer Science

Keywords

Categorization
Object Recognition

Access to Document

10.1007/978-3-319-10590-1_4

Cite this

Deng, J., Ding, N., Jia, Y., Frome, A., Murphy, K., Bengio, S., Li, Y., Neven, H., & Adam, H. (2014). Large-scale object classification using label relation graphs. In Computer Vision, ECCV 2014 - 13th European Conference, Proceedings (PART 1 ed., pp. 48-64). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8689 LNCS, No. PART 1). Springer Verlag. https://doi.org/10.1007/978-3-319-10590-1_4

@inproceedings{2e31e9376d0e43cf8166c3c0676b3df7,

title = "Large-scale object classification using label relation graphs",

abstract = "In this paper we study how to perform object classification in a principled way that exploits the rich structure of real world labels. We develop a new model that allows encoding of flexible relations between labels. We introduce Hierarchy and Exclusion (HEX) graphs, a new formalism that captures semantic relations between any two labels applied to the same object: mutual exclusion, overlap and subsumption. We then provide rigorous theoretical analysis that illustrates properties of HEX graphs such as consistency, equivalence, and computational implications of the graph structure. Next, we propose a probabilistic classification model based on HEX graphs and show that it enjoys a number of desirable properties. Finally, we evaluate our method using a large-scale benchmark. Empirical results demonstrate that our model can significantly improve object classification by exploiting the label relations.",

keywords = "Categorization, Object Recognition",

author = "Jia Deng and Nan Ding and Yangqing Jia and Andrea Frome and Kevin Murphy and Samy Bengio and Yuan Li and Hartmut Neven and Hartwig Adam",

year = "2014",

doi = "10.1007/978-3-319-10590-1_4",

language = "English (US)",

isbn = "9783319105895",

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

publisher = "Springer Verlag",

number = "PART 1",

pages = "48--64",

booktitle = "Computer Vision, ECCV 2014 - 13th European Conference, Proceedings",

address = "Germany",

edition = "PART 1",

note = "13th European Conference on Computer Vision, ECCV 2014 ; Conference date: 06-09-2014 Through 12-09-2014",

}

Deng, J, Ding, N, Jia, Y, Frome, A, Murphy, K, Bengio, S, Li, Y, Neven, H & Adam, H 2014, Large-scale object classification using label relation graphs. in Computer Vision, ECCV 2014 - 13th European Conference, Proceedings. PART 1 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 1, vol. 8689 LNCS, Springer Verlag, pp. 48-64, 13th European Conference on Computer Vision, ECCV 2014, Zurich, Switzerland, 9/6/14. https://doi.org/10.1007/978-3-319-10590-1_4

Large-scale object classification using label relation graphs. / Deng, Jia; Ding, Nan; Jia, Yangqing et al.
Computer Vision, ECCV 2014 - 13th European Conference, Proceedings. PART 1. ed. Springer Verlag, 2014. p. 48-64 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 8689 LNCS, No. PART 1).

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

TY - GEN

T1 - Large-scale object classification using label relation graphs

AU - Deng, Jia

AU - Ding, Nan

AU - Jia, Yangqing

AU - Frome, Andrea

AU - Murphy, Kevin

AU - Bengio, Samy

AU - Li, Yuan

AU - Neven, Hartmut

AU - Adam, Hartwig

PY - 2014

Y1 - 2014

N2 - In this paper we study how to perform object classification in a principled way that exploits the rich structure of real world labels. We develop a new model that allows encoding of flexible relations between labels. We introduce Hierarchy and Exclusion (HEX) graphs, a new formalism that captures semantic relations between any two labels applied to the same object: mutual exclusion, overlap and subsumption. We then provide rigorous theoretical analysis that illustrates properties of HEX graphs such as consistency, equivalence, and computational implications of the graph structure. Next, we propose a probabilistic classification model based on HEX graphs and show that it enjoys a number of desirable properties. Finally, we evaluate our method using a large-scale benchmark. Empirical results demonstrate that our model can significantly improve object classification by exploiting the label relations.

AB - In this paper we study how to perform object classification in a principled way that exploits the rich structure of real world labels. We develop a new model that allows encoding of flexible relations between labels. We introduce Hierarchy and Exclusion (HEX) graphs, a new formalism that captures semantic relations between any two labels applied to the same object: mutual exclusion, overlap and subsumption. We then provide rigorous theoretical analysis that illustrates properties of HEX graphs such as consistency, equivalence, and computational implications of the graph structure. Next, we propose a probabilistic classification model based on HEX graphs and show that it enjoys a number of desirable properties. Finally, we evaluate our method using a large-scale benchmark. Empirical results demonstrate that our model can significantly improve object classification by exploiting the label relations.

KW - Categorization

KW - Object Recognition

UR - http://www.scopus.com/inward/record.url?scp=84906513601&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84906513601&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-10590-1_4

DO - 10.1007/978-3-319-10590-1_4

M3 - Conference contribution

AN - SCOPUS:84906513601

SN - 9783319105895

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

SP - 48

EP - 64

BT - Computer Vision, ECCV 2014 - 13th European Conference, Proceedings

PB - Springer Verlag

T2 - 13th European Conference on Computer Vision, ECCV 2014

Y2 - 6 September 2014 through 12 September 2014

ER -

Deng J, Ding N, Jia Y, Frome A, Murphy K, Bengio S et al. Large-scale object classification using label relation graphs. In Computer Vision, ECCV 2014 - 13th European Conference, Proceedings. PART 1 ed. Springer Verlag. 2014. p. 48-64. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1). doi: 10.1007/978-3-319-10590-1_4

Large-scale object classification using label relation graphs

Abstract

Publication series

Other

All Science Journal Classification (ASJC) codes

Keywords

Access to Document

Other files and links

Fingerprint

Cite this