Bootstrap dual complementary hashing with semi-supervised re-ranking for image retrieval

Xing Tian, Xiancheng Zhou, Wing Ng, Jiayong Li, Hui Wang

Research output: Contribution to journalArticle

Abstract

With the rapid growth of multimedia data on the Internet, content-based image retrieval becomes a key technique for the Internet development. Hashing methods are efficient and effective for image retrieval. Dual Complementary Hashing (DCH) is one such method, which uses multiple hash tables and has good performance. However, DCH utilizes wrongly hashed image pairs to train the following hash table and discards correctly hashed image pairs. Therefore, the number of image pairs utilized for training the following hash tables will decrease rapidly. Moreover, each hash function in a hash table of DCH is trained by correcting the errors caused by its preceding one instead of holistically considering errors made by all previous hash functions. These restrictions significantly reduce the training efficiency and the overall performance of DCH. In this paper, we propose a new hashing method for image retrieval, Bootstrap Dual Complementary Hashing with semi-supervised Re-ranking (BDCHR). It is a semi-supervised multi-hashing method consisting of two parts: bootstrap DCH and semi-supervised re-ranking. The first part relieves the restrictions of DCH while the second part further enhances the image retrieval performance. Experimental results show that BDCHR yields better performance than other state-of-the-art multi-hashing methods.

Original languageEnglish
Pages (from-to)103-116
Number of pages14
JournalNeurocomputing
Volume379
DOIs
Publication statusPublished - 28 Feb 2020

Fingerprint

Image retrieval
Hash functions
Internet
Multimedia
Growth

Keywords

  • Multi-hashing
  • dual complementary hashing
  • image retrieval
  • semi-supervised

Cite this

Tian, Xing ; Zhou, Xiancheng ; Ng, Wing ; Li, Jiayong ; Wang, Hui. / Bootstrap dual complementary hashing with semi-supervised re-ranking for image retrieval. In: Neurocomputing. 2020 ; Vol. 379. pp. 103-116.
@article{2a66ede7adac49cca9243006bdb072d8,
title = "Bootstrap dual complementary hashing with semi-supervised re-ranking for image retrieval",
abstract = "With the rapid growth of multimedia data on the Internet, content-based image retrieval becomes a key technique for the Internet development. Hashing methods are efficient and effective for image retrieval. Dual Complementary Hashing (DCH) is one such method, which uses multiple hash tables and has good performance. However, DCH utilizes wrongly hashed image pairs to train the following hash table and discards correctly hashed image pairs. Therefore, the number of image pairs utilized for training the following hash tables will decrease rapidly. Moreover, each hash function in a hash table of DCH is trained by correcting the errors caused by its preceding one instead of holistically considering errors made by all previous hash functions. These restrictions significantly reduce the training efficiency and the overall performance of DCH. In this paper, we propose a new hashing method for image retrieval, Bootstrap Dual Complementary Hashing with semi-supervised Re-ranking (BDCHR). It is a semi-supervised multi-hashing method consisting of two parts: bootstrap DCH and semi-supervised re-ranking. The first part relieves the restrictions of DCH while the second part further enhances the image retrieval performance. Experimental results show that BDCHR yields better performance than other state-of-the-art multi-hashing methods.",
keywords = "Multi-hashing, dual complementary hashing, image retrieval, semi-supervised",
author = "Xing Tian and Xiancheng Zhou and Wing Ng and Jiayong Li and Hui Wang",
year = "2020",
month = "2",
day = "28",
doi = "10.1016/j.neucom.2019.10.073",
language = "English",
volume = "379",
pages = "103--116",
journal = "Neurocomputing",
issn = "0925-2312",
publisher = "Elsevier",

}

Bootstrap dual complementary hashing with semi-supervised re-ranking for image retrieval. / Tian, Xing; Zhou, Xiancheng; Ng, Wing; Li, Jiayong; Wang, Hui.

In: Neurocomputing, Vol. 379, 28.02.2020, p. 103-116.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Bootstrap dual complementary hashing with semi-supervised re-ranking for image retrieval

AU - Tian, Xing

AU - Zhou, Xiancheng

AU - Ng, Wing

AU - Li, Jiayong

AU - Wang, Hui

PY - 2020/2/28

Y1 - 2020/2/28

N2 - With the rapid growth of multimedia data on the Internet, content-based image retrieval becomes a key technique for the Internet development. Hashing methods are efficient and effective for image retrieval. Dual Complementary Hashing (DCH) is one such method, which uses multiple hash tables and has good performance. However, DCH utilizes wrongly hashed image pairs to train the following hash table and discards correctly hashed image pairs. Therefore, the number of image pairs utilized for training the following hash tables will decrease rapidly. Moreover, each hash function in a hash table of DCH is trained by correcting the errors caused by its preceding one instead of holistically considering errors made by all previous hash functions. These restrictions significantly reduce the training efficiency and the overall performance of DCH. In this paper, we propose a new hashing method for image retrieval, Bootstrap Dual Complementary Hashing with semi-supervised Re-ranking (BDCHR). It is a semi-supervised multi-hashing method consisting of two parts: bootstrap DCH and semi-supervised re-ranking. The first part relieves the restrictions of DCH while the second part further enhances the image retrieval performance. Experimental results show that BDCHR yields better performance than other state-of-the-art multi-hashing methods.

AB - With the rapid growth of multimedia data on the Internet, content-based image retrieval becomes a key technique for the Internet development. Hashing methods are efficient and effective for image retrieval. Dual Complementary Hashing (DCH) is one such method, which uses multiple hash tables and has good performance. However, DCH utilizes wrongly hashed image pairs to train the following hash table and discards correctly hashed image pairs. Therefore, the number of image pairs utilized for training the following hash tables will decrease rapidly. Moreover, each hash function in a hash table of DCH is trained by correcting the errors caused by its preceding one instead of holistically considering errors made by all previous hash functions. These restrictions significantly reduce the training efficiency and the overall performance of DCH. In this paper, we propose a new hashing method for image retrieval, Bootstrap Dual Complementary Hashing with semi-supervised Re-ranking (BDCHR). It is a semi-supervised multi-hashing method consisting of two parts: bootstrap DCH and semi-supervised re-ranking. The first part relieves the restrictions of DCH while the second part further enhances the image retrieval performance. Experimental results show that BDCHR yields better performance than other state-of-the-art multi-hashing methods.

KW - Multi-hashing

KW - dual complementary hashing

KW - image retrieval

KW - semi-supervised

UR - http://www.scopus.com/inward/record.url?scp=85075422460&partnerID=8YFLogxK

U2 - 10.1016/j.neucom.2019.10.073

DO - 10.1016/j.neucom.2019.10.073

M3 - Article

VL - 379

SP - 103

EP - 116

JO - Neurocomputing

JF - Neurocomputing

SN - 0925-2312

ER -