Unsupervised Topic Hypergraph Hashing for Efficient Mobile Image Retrieval

Lei Zhu, Jialie Shen, Liang Xie, Zhiyong Cheng

Research output: Contribution to journalArticleResearchpeer-review

Abstract

Hashing compresses high-dimensional features into compact binary codes. It is one of the promising techniques to support efficient mobile image retrieval, due to its low data transmission cost and fast retrieval response. However, most of existing hashing strategies simply rely on low-level features. Thus, they may generate hashing codes with limited discriminative capability. Moreover, many of them fail to exploit complex and high-order semantic correlations that inherently exist among images. Motivated by these observations, we propose a novel unsupervised hashing scheme, called topic hypergraph hashing (THH), to address the limitations. THH effectively mitigates the semantic shortage of hashing codes by exploiting auxiliary texts around images. In our method, relations between images and semantic topics are first discovered via robust collective non-negative matrix factorization. Afterwards, a unified topic hypergraph, where images and topics are represented with independent vertices and hyperedges, respectively, is constructed to model inherent high-order semantic correlations of images. Finally, hashing codes and functions are learned by simultaneously enforcing semantic consistence and preserving the discovered semantic relations. Experiments on publicly available datasets demonstrate that THH can achieve superior performance compared with several state-of-The-Art methods, and it is more suitable for mobile image retrieval.

Original languageEnglish
Article number17256518
Pages (from-to)3941-3954
Number of pages14
JournalIEEE Transactions on Cybernetics
Volume47
Issue number11
DOIs
Publication statusPublished - 1 Nov 2017
Externally publishedYes

Fingerprint

Image retrieval
Semantics
Binary codes
Factorization
Data communication systems
Costs
Experiments

Cite this

Zhu, Lei ; Shen, Jialie ; Xie, Liang ; Cheng, Zhiyong. / Unsupervised Topic Hypergraph Hashing for Efficient Mobile Image Retrieval. In: IEEE Transactions on Cybernetics. 2017 ; Vol. 47, No. 11. pp. 3941-3954.
@article{9b53b63c83254c238b49262ac32867b3,
title = "Unsupervised Topic Hypergraph Hashing for Efficient Mobile Image Retrieval",
abstract = "Hashing compresses high-dimensional features into compact binary codes. It is one of the promising techniques to support efficient mobile image retrieval, due to its low data transmission cost and fast retrieval response. However, most of existing hashing strategies simply rely on low-level features. Thus, they may generate hashing codes with limited discriminative capability. Moreover, many of them fail to exploit complex and high-order semantic correlations that inherently exist among images. Motivated by these observations, we propose a novel unsupervised hashing scheme, called topic hypergraph hashing (THH), to address the limitations. THH effectively mitigates the semantic shortage of hashing codes by exploiting auxiliary texts around images. In our method, relations between images and semantic topics are first discovered via robust collective non-negative matrix factorization. Afterwards, a unified topic hypergraph, where images and topics are represented with independent vertices and hyperedges, respectively, is constructed to model inherent high-order semantic correlations of images. Finally, hashing codes and functions are learned by simultaneously enforcing semantic consistence and preserving the discovered semantic relations. Experiments on publicly available datasets demonstrate that THH can achieve superior performance compared with several state-of-The-Art methods, and it is more suitable for mobile image retrieval.",
keywords = "High-order semantic correlations, mobile image retrieval, topic hypergraph hashing (THH)",
author = "Lei Zhu and Jialie Shen and Liang Xie and Zhiyong Cheng",
year = "2017",
month = "11",
day = "1",
doi = "10.1109/TCYB.2016.2591068",
language = "English",
volume = "47",
pages = "3941--3954",
journal = "IEEE Transactions on Cybernetics",
issn = "1083-4419",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",
number = "11",

}

Unsupervised Topic Hypergraph Hashing for Efficient Mobile Image Retrieval. / Zhu, Lei; Shen, Jialie; Xie, Liang; Cheng, Zhiyong.

In: IEEE Transactions on Cybernetics, Vol. 47, No. 11, 17256518, 01.11.2017, p. 3941-3954.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Unsupervised Topic Hypergraph Hashing for Efficient Mobile Image Retrieval

AU - Zhu, Lei

AU - Shen, Jialie

AU - Xie, Liang

AU - Cheng, Zhiyong

PY - 2017/11/1

Y1 - 2017/11/1

N2 - Hashing compresses high-dimensional features into compact binary codes. It is one of the promising techniques to support efficient mobile image retrieval, due to its low data transmission cost and fast retrieval response. However, most of existing hashing strategies simply rely on low-level features. Thus, they may generate hashing codes with limited discriminative capability. Moreover, many of them fail to exploit complex and high-order semantic correlations that inherently exist among images. Motivated by these observations, we propose a novel unsupervised hashing scheme, called topic hypergraph hashing (THH), to address the limitations. THH effectively mitigates the semantic shortage of hashing codes by exploiting auxiliary texts around images. In our method, relations between images and semantic topics are first discovered via robust collective non-negative matrix factorization. Afterwards, a unified topic hypergraph, where images and topics are represented with independent vertices and hyperedges, respectively, is constructed to model inherent high-order semantic correlations of images. Finally, hashing codes and functions are learned by simultaneously enforcing semantic consistence and preserving the discovered semantic relations. Experiments on publicly available datasets demonstrate that THH can achieve superior performance compared with several state-of-The-Art methods, and it is more suitable for mobile image retrieval.

AB - Hashing compresses high-dimensional features into compact binary codes. It is one of the promising techniques to support efficient mobile image retrieval, due to its low data transmission cost and fast retrieval response. However, most of existing hashing strategies simply rely on low-level features. Thus, they may generate hashing codes with limited discriminative capability. Moreover, many of them fail to exploit complex and high-order semantic correlations that inherently exist among images. Motivated by these observations, we propose a novel unsupervised hashing scheme, called topic hypergraph hashing (THH), to address the limitations. THH effectively mitigates the semantic shortage of hashing codes by exploiting auxiliary texts around images. In our method, relations between images and semantic topics are first discovered via robust collective non-negative matrix factorization. Afterwards, a unified topic hypergraph, where images and topics are represented with independent vertices and hyperedges, respectively, is constructed to model inherent high-order semantic correlations of images. Finally, hashing codes and functions are learned by simultaneously enforcing semantic consistence and preserving the discovered semantic relations. Experiments on publicly available datasets demonstrate that THH can achieve superior performance compared with several state-of-The-Art methods, and it is more suitable for mobile image retrieval.

KW - High-order semantic correlations

KW - mobile image retrieval

KW - topic hypergraph hashing (THH)

UR - http://www.scopus.com/inward/record.url?scp=84992152145&partnerID=8YFLogxK

U2 - 10.1109/TCYB.2016.2591068

DO - 10.1109/TCYB.2016.2591068

M3 - Article

VL - 47

SP - 3941

EP - 3954

JO - IEEE Transactions on Cybernetics

JF - IEEE Transactions on Cybernetics

SN - 1083-4419

IS - 11

M1 - 17256518

ER -