Social tag relevance learning via ranking-oriented neighbor voting

Chaoran Cui, Jialie Shen, Jun Ma, Tao Lian

Research output: Contribution to journalArticleResearchpeer-review

Abstract

High quality tags play a critical role in applications involving online multimedia search, such as social image annotation, sharing and browsing. However, user-generated tags in real world are often imprecise and incomplete to describe the image contents, which severely degrades the performance of current search systems. To improve the descriptive powers of social tags, a fundamental issue is tag relevance learning, which concerns how to interpret the relevance of a tag with respect to the contents of an image effectively. In this paper, we investigate the problem from a new perspective of learning to rank, and develop a novel approach to facilitate tag relevance learning to directly optimize the ranking performance of tag-based image search. Specifically, a supervision step is introduced into the neighbor voting scheme, in which the tag relevance is estimated by accumulating votes from visual neighbors. Through explicitly modeling the neighbor weights and tag correlations, the risk of making heuristic assumptions is effectively avoided. Besides, our approach does not suffer from the scalability problem since a generic model is learned that can be applied to all tags. Extensive experiments on two benchmark datasets in comparison with the state-of-the-art methods demonstrate the promise of our approach.

Original languageEnglish
Pages (from-to)8831-8857
Number of pages27
JournalMultimedia Tools and Applications
Volume76
Issue number6
DOIs
Publication statusPublished - Mar 2017
Externally publishedYes

Fingerprint

Scalability
Experiments

Cite this

Cui, Chaoran ; Shen, Jialie ; Ma, Jun ; Lian, Tao. / Social tag relevance learning via ranking-oriented neighbor voting. In: Multimedia Tools and Applications. 2017 ; Vol. 76, No. 6. pp. 8831-8857.
@article{38670d32a48f4e04beac3463123a08af,
title = "Social tag relevance learning via ranking-oriented neighbor voting",
abstract = "High quality tags play a critical role in applications involving online multimedia search, such as social image annotation, sharing and browsing. However, user-generated tags in real world are often imprecise and incomplete to describe the image contents, which severely degrades the performance of current search systems. To improve the descriptive powers of social tags, a fundamental issue is tag relevance learning, which concerns how to interpret the relevance of a tag with respect to the contents of an image effectively. In this paper, we investigate the problem from a new perspective of learning to rank, and develop a novel approach to facilitate tag relevance learning to directly optimize the ranking performance of tag-based image search. Specifically, a supervision step is introduced into the neighbor voting scheme, in which the tag relevance is estimated by accumulating votes from visual neighbors. Through explicitly modeling the neighbor weights and tag correlations, the risk of making heuristic assumptions is effectively avoided. Besides, our approach does not suffer from the scalability problem since a generic model is learned that can be applied to all tags. Extensive experiments on two benchmark datasets in comparison with the state-of-the-art methods demonstrate the promise of our approach.",
keywords = "Learning to rank, Neighbor voting, Tag relevance learning, Tag-based image search",
author = "Chaoran Cui and Jialie Shen and Jun Ma and Tao Lian",
year = "2017",
month = "3",
doi = "10.1007/s11042-016-3512-1",
language = "English",
volume = "76",
pages = "8831--8857",
journal = "Multimedia Tools and Applications",
issn = "1380-7501",
publisher = "Springer Netherlands",
number = "6",

}

Social tag relevance learning via ranking-oriented neighbor voting. / Cui, Chaoran; Shen, Jialie; Ma, Jun; Lian, Tao.

In: Multimedia Tools and Applications, Vol. 76, No. 6, 03.2017, p. 8831-8857.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Social tag relevance learning via ranking-oriented neighbor voting

AU - Cui, Chaoran

AU - Shen, Jialie

AU - Ma, Jun

AU - Lian, Tao

PY - 2017/3

Y1 - 2017/3

N2 - High quality tags play a critical role in applications involving online multimedia search, such as social image annotation, sharing and browsing. However, user-generated tags in real world are often imprecise and incomplete to describe the image contents, which severely degrades the performance of current search systems. To improve the descriptive powers of social tags, a fundamental issue is tag relevance learning, which concerns how to interpret the relevance of a tag with respect to the contents of an image effectively. In this paper, we investigate the problem from a new perspective of learning to rank, and develop a novel approach to facilitate tag relevance learning to directly optimize the ranking performance of tag-based image search. Specifically, a supervision step is introduced into the neighbor voting scheme, in which the tag relevance is estimated by accumulating votes from visual neighbors. Through explicitly modeling the neighbor weights and tag correlations, the risk of making heuristic assumptions is effectively avoided. Besides, our approach does not suffer from the scalability problem since a generic model is learned that can be applied to all tags. Extensive experiments on two benchmark datasets in comparison with the state-of-the-art methods demonstrate the promise of our approach.

AB - High quality tags play a critical role in applications involving online multimedia search, such as social image annotation, sharing and browsing. However, user-generated tags in real world are often imprecise and incomplete to describe the image contents, which severely degrades the performance of current search systems. To improve the descriptive powers of social tags, a fundamental issue is tag relevance learning, which concerns how to interpret the relevance of a tag with respect to the contents of an image effectively. In this paper, we investigate the problem from a new perspective of learning to rank, and develop a novel approach to facilitate tag relevance learning to directly optimize the ranking performance of tag-based image search. Specifically, a supervision step is introduced into the neighbor voting scheme, in which the tag relevance is estimated by accumulating votes from visual neighbors. Through explicitly modeling the neighbor weights and tag correlations, the risk of making heuristic assumptions is effectively avoided. Besides, our approach does not suffer from the scalability problem since a generic model is learned that can be applied to all tags. Extensive experiments on two benchmark datasets in comparison with the state-of-the-art methods demonstrate the promise of our approach.

KW - Learning to rank

KW - Neighbor voting

KW - Tag relevance learning

KW - Tag-based image search

UR - http://www.scopus.com/inward/record.url?scp=84963706207&partnerID=8YFLogxK

U2 - 10.1007/s11042-016-3512-1

DO - 10.1007/s11042-016-3512-1

M3 - Article

VL - 76

SP - 8831

EP - 8857

JO - Multimedia Tools and Applications

JF - Multimedia Tools and Applications

SN - 1380-7501

IS - 6

ER -