Locality sensitive hashing for content based image retrieval: A comparative experimental study

Sanaa Chafik; Imane Daoudi; Hamid El Ouardi; Mounim A.El Yacoubi; Bernadette Dorizzi

doi:10.1109/NGNS.2014.6990224

Locality sensitive hashing for content based image retrieval: A comparative experimental study

Sanaa Chafik, Imane Daoudi, Hamid El Ouardi, Mounim A.El Yacoubi, Bernadette Dorizzi

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

Abstract

This paper presents a comparative experimental study of the multidimensional indexing methods based on the approximation approach. We are particularly interested in the LSH family, which provides efficient index structures and solves the dimensionality curse problem. The goal is to understand the performance gain and the behavior of this family of methods on large-scale databases. E2LSH is compared to the KRA+-Blocks and the sequential scan methods. Two criteria are used in evaluating the E2LSH performances, namely average precision and CPU time using a database of one million image descriptors.

Original language	English
Title of host publication	International Conference on Next Generation Networks and Services, NGNS
Publisher	IEEE Computer Society
Pages	38-43
Number of pages	6
ISBN (Electronic)	9781479969371
DOIs	https://doi.org/10.1109/NGNS.2014.6990224
Publication status	Published - 16 Dec 2014
Externally published	Yes
Event	2014 5th International Conference on Next Generation Networks and Services, NGNS 2014 - Casablanca, Morocco Duration: 28 May 2014 → 30 May 2014

Publication series

Name	International Conference on Next Generation Networks and Services, NGNS
ISSN (Print)	2327-6525
ISSN (Electronic)	2327-6533

Conference

Conference	2014 5th International Conference on Next Generation Networks and Services, NGNS 2014
Country/Territory	Morocco
City	Casablanca
Period	28/05/14 → 30/05/14

Keywords

Content based image retrieval (CBIR)
Curse of dimensionality
Locality sensitive hashing
Multidimensional indexing
Scalability

Access to Document

10.1109/NGNS.2014.6990224

Cite this

Chafik, S., Daoudi, I., Ouardi, H. E., Yacoubi, M. A. E., & Dorizzi, B. (2014). Locality sensitive hashing for content based image retrieval: A comparative experimental study. In International Conference on Next Generation Networks and Services, NGNS (pp. 38-43). Article 6990224 (International Conference on Next Generation Networks and Services, NGNS). IEEE Computer Society. https://doi.org/10.1109/NGNS.2014.6990224

@inproceedings{b902732abc2b44eea8ebe79e99523743,

title = "Locality sensitive hashing for content based image retrieval: A comparative experimental study",

abstract = "This paper presents a comparative experimental study of the multidimensional indexing methods based on the approximation approach. We are particularly interested in the LSH family, which provides efficient index structures and solves the dimensionality curse problem. The goal is to understand the performance gain and the behavior of this family of methods on large-scale databases. E2LSH is compared to the KRA+-Blocks and the sequential scan methods. Two criteria are used in evaluating the E2LSH performances, namely average precision and CPU time using a database of one million image descriptors.",

keywords = "Content based image retrieval (CBIR), Curse of dimensionality, Locality sensitive hashing, Multidimensional indexing, Scalability",

author = "Sanaa Chafik and Imane Daoudi and Ouardi, \{Hamid El\} and Yacoubi, \{Mounim A.El\} and Bernadette Dorizzi",

note = "Publisher Copyright: {\textcopyright} 2014 IEEE.; 2014 5th International Conference on Next Generation Networks and Services, NGNS 2014 ; Conference date: 28-05-2014 Through 30-05-2014",

year = "2014",

month = dec,

day = "16",

doi = "10.1109/NGNS.2014.6990224",

language = "English",

series = "International Conference on Next Generation Networks and Services, NGNS",

publisher = "IEEE Computer Society",

pages = "38--43",

booktitle = "International Conference on Next Generation Networks and Services, NGNS",

}

Chafik, S, Daoudi, I, Ouardi, HE, Yacoubi, MAE & Dorizzi, B 2014, Locality sensitive hashing for content based image retrieval: A comparative experimental study. in International Conference on Next Generation Networks and Services, NGNS., 6990224, International Conference on Next Generation Networks and Services, NGNS, IEEE Computer Society, pp. 38-43, 2014 5th International Conference on Next Generation Networks and Services, NGNS 2014, Casablanca, Morocco, 28/05/14. https://doi.org/10.1109/NGNS.2014.6990224

Locality sensitive hashing for content based image retrieval: A comparative experimental study. / Chafik, Sanaa; Daoudi, Imane; Ouardi, Hamid El et al.
International Conference on Next Generation Networks and Services, NGNS. IEEE Computer Society, 2014. p. 38-43 6990224 (International Conference on Next Generation Networks and Services, NGNS).

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

TY - GEN

T1 - Locality sensitive hashing for content based image retrieval

T2 - 2014 5th International Conference on Next Generation Networks and Services, NGNS 2014

AU - Chafik, Sanaa

AU - Daoudi, Imane

AU - Ouardi, Hamid El

AU - Yacoubi, Mounim A.El

AU - Dorizzi, Bernadette

PY - 2014/12/16

Y1 - 2014/12/16

N2 - This paper presents a comparative experimental study of the multidimensional indexing methods based on the approximation approach. We are particularly interested in the LSH family, which provides efficient index structures and solves the dimensionality curse problem. The goal is to understand the performance gain and the behavior of this family of methods on large-scale databases. E2LSH is compared to the KRA+-Blocks and the sequential scan methods. Two criteria are used in evaluating the E2LSH performances, namely average precision and CPU time using a database of one million image descriptors.

AB - This paper presents a comparative experimental study of the multidimensional indexing methods based on the approximation approach. We are particularly interested in the LSH family, which provides efficient index structures and solves the dimensionality curse problem. The goal is to understand the performance gain and the behavior of this family of methods on large-scale databases. E2LSH is compared to the KRA+-Blocks and the sequential scan methods. Two criteria are used in evaluating the E2LSH performances, namely average precision and CPU time using a database of one million image descriptors.

KW - Content based image retrieval (CBIR)

KW - Curse of dimensionality

KW - Locality sensitive hashing

KW - Multidimensional indexing

KW - Scalability

U2 - 10.1109/NGNS.2014.6990224

DO - 10.1109/NGNS.2014.6990224

M3 - Conference contribution

AN - SCOPUS:84920013102

T3 - International Conference on Next Generation Networks and Services, NGNS

SP - 38

EP - 43

BT - International Conference on Next Generation Networks and Services, NGNS

PB - IEEE Computer Society

Y2 - 28 May 2014 through 30 May 2014

ER -

Chafik S, Daoudi I, Ouardi HE, Yacoubi MAE, Dorizzi B. Locality sensitive hashing for content based image retrieval: A comparative experimental study. In International Conference on Next Generation Networks and Services, NGNS. IEEE Computer Society. 2014. p. 38-43. 6990224. (International Conference on Next Generation Networks and Services, NGNS). doi: 10.1109/NGNS.2014.6990224

Locality sensitive hashing for content based image retrieval: A comparative experimental study

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this