Randomized locality sensitive vocabularies for bag-of-features model

Yadong Mu; Ju Sun; Tony X. Han; Loong Fah Cheong; Shuicheng Yan

doi:10.1007/978-3-642-15558-1_54

Randomized locality sensitive vocabularies for bag-of-features model

Yadong Mu, Ju Sun, Tony X. Han, Loong Fah Cheong, Shuicheng Yan

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

17 Scopus citations

Abstract

Visual vocabulary construction is an integral part of the popular Bag-of-Features (BOF) model. When visual data scale up (in terms of the dimensionality of features or/and the number of samples), most existing algorithms (e.g. k-means) become unfavorable due to the prohibitive time and space requirements. In this paper we propose the random locality sensitive vocabulary (RLSV) scheme towards efficient visual vocabulary construction in such scenarios. Integrating ideas from the Locality Sensitive Hashing (LSH) and the Random Forest (RF), RLSV generates and aggregates multiple visual vocabularies based on random projections, without taking clustering or training efforts. This simple scheme demonstrates superior time and space efficiency over prior methods, in both theory and practice, while often achieving comparable or even better performances. Besides, extensions to supervised and kernelized vocabulary constructions are also discussed and experimented with.

Original language	English (US)
Title of host publication	Computer Vision, ECCV 2010 - 11th European Conference on Computer Vision, Proceedings
Publisher	Springer Verlag
Pages	748-761
Number of pages	14
Edition	PART 3
ISBN (Print)	364215557X, 9783642155574
DOIs	https://doi.org/10.1007/978-3-642-15558-1_54
State	Published - Jan 1 2010
Externally published	Yes
Event	11th European Conference on Computer Vision, ECCV 2010 - Heraklion, Crete, Greece Duration: Sep 10 2010 → Sep 11 2010

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 3
Volume	6313 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	11th European Conference on Computer Vision, ECCV 2010
Country	Greece
City	Heraklion, Crete
Period	9/10/10 → 9/11/10

Access

10.1007/978-3-642-15558-1_54

Fingerprint Dive into the research topics of 'Randomized locality sensitive vocabularies for bag-of-features model'. Together they form a unique fingerprint.

Cite this

Mu, Y., Sun, J., Han, T. X., Cheong, L. F., & Yan, S. (2010). Randomized locality sensitive vocabularies for bag-of-features model. In Computer Vision, ECCV 2010 - 11th European Conference on Computer Vision, Proceedings (PART 3 ed., pp. 748-761). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6313 LNCS, No. PART 3). Springer Verlag. https://doi.org/10.1007/978-3-642-15558-1_54

Randomized locality sensitive vocabularies for bag-of-features model. / Mu, Yadong; Sun, Ju; Han, Tony X.; Cheong, Loong Fah; Yan, Shuicheng.

Computer Vision, ECCV 2010 - 11th European Conference on Computer Vision, Proceedings. PART 3. ed. Springer Verlag, 2010. p. 748-761 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6313 LNCS, No. PART 3).

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

Mu, Y, Sun, J, Han, TX, Cheong, LF & Yan, S 2010, Randomized locality sensitive vocabularies for bag-of-features model. in Computer Vision, ECCV 2010 - 11th European Conference on Computer Vision, Proceedings. PART 3 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 3, vol. 6313 LNCS, Springer Verlag, pp. 748-761, 11th European Conference on Computer Vision, ECCV 2010, Heraklion, Crete, Greece, 9/10/10. https://doi.org/10.1007/978-3-642-15558-1_54

Mu Y, Sun J, Han TX, Cheong LF, Yan S. Randomized locality sensitive vocabularies for bag-of-features model. In Computer Vision, ECCV 2010 - 11th European Conference on Computer Vision, Proceedings. PART 3 ed. Springer Verlag. 2010. p. 748-761. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 3). https://doi.org/10.1007/978-3-642-15558-1_54

Mu, Yadong ; Sun, Ju ; Han, Tony X. ; Cheong, Loong Fah ; Yan, Shuicheng. / Randomized locality sensitive vocabularies for bag-of-features model. Computer Vision, ECCV 2010 - 11th European Conference on Computer Vision, Proceedings. PART 3. ed. Springer Verlag, 2010. pp. 748-761 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 3).

@inproceedings{9068075fa6374f8e8ade664cf27458c4,

title = "Randomized locality sensitive vocabularies for bag-of-features model",

abstract = "Visual vocabulary construction is an integral part of the popular Bag-of-Features (BOF) model. When visual data scale up (in terms of the dimensionality of features or/and the number of samples), most existing algorithms (e.g. k-means) become unfavorable due to the prohibitive time and space requirements. In this paper we propose the random locality sensitive vocabulary (RLSV) scheme towards efficient visual vocabulary construction in such scenarios. Integrating ideas from the Locality Sensitive Hashing (LSH) and the Random Forest (RF), RLSV generates and aggregates multiple visual vocabularies based on random projections, without taking clustering or training efforts. This simple scheme demonstrates superior time and space efficiency over prior methods, in both theory and practice, while often achieving comparable or even better performances. Besides, extensions to supervised and kernelized vocabulary constructions are also discussed and experimented with.",

author = "Yadong Mu and Ju Sun and Han, {Tony X.} and Cheong, {Loong Fah} and Shuicheng Yan",

year = "2010",

month = jan,

day = "1",

doi = "10.1007/978-3-642-15558-1_54",

language = "English (US)",

isbn = "364215557X",

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

publisher = "Springer Verlag",

number = "PART 3",

pages = "748--761",

booktitle = "Computer Vision, ECCV 2010 - 11th European Conference on Computer Vision, Proceedings",

edition = "PART 3",

note = "11th European Conference on Computer Vision, ECCV 2010 ; Conference date: 10-09-2010 Through 11-09-2010",

}

TY - GEN

T1 - Randomized locality sensitive vocabularies for bag-of-features model

AU - Mu, Yadong

AU - Sun, Ju

AU - Han, Tony X.

AU - Cheong, Loong Fah

AU - Yan, Shuicheng

PY - 2010/1/1

Y1 - 2010/1/1

N2 - Visual vocabulary construction is an integral part of the popular Bag-of-Features (BOF) model. When visual data scale up (in terms of the dimensionality of features or/and the number of samples), most existing algorithms (e.g. k-means) become unfavorable due to the prohibitive time and space requirements. In this paper we propose the random locality sensitive vocabulary (RLSV) scheme towards efficient visual vocabulary construction in such scenarios. Integrating ideas from the Locality Sensitive Hashing (LSH) and the Random Forest (RF), RLSV generates and aggregates multiple visual vocabularies based on random projections, without taking clustering or training efforts. This simple scheme demonstrates superior time and space efficiency over prior methods, in both theory and practice, while often achieving comparable or even better performances. Besides, extensions to supervised and kernelized vocabulary constructions are also discussed and experimented with.

AB - Visual vocabulary construction is an integral part of the popular Bag-of-Features (BOF) model. When visual data scale up (in terms of the dimensionality of features or/and the number of samples), most existing algorithms (e.g. k-means) become unfavorable due to the prohibitive time and space requirements. In this paper we propose the random locality sensitive vocabulary (RLSV) scheme towards efficient visual vocabulary construction in such scenarios. Integrating ideas from the Locality Sensitive Hashing (LSH) and the Random Forest (RF), RLSV generates and aggregates multiple visual vocabularies based on random projections, without taking clustering or training efforts. This simple scheme demonstrates superior time and space efficiency over prior methods, in both theory and practice, while often achieving comparable or even better performances. Besides, extensions to supervised and kernelized vocabulary constructions are also discussed and experimented with.

UR - http://www.scopus.com/inward/record.url?scp=78149312744&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=78149312744&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-15558-1_54

DO - 10.1007/978-3-642-15558-1_54

M3 - Conference contribution

AN - SCOPUS:78149312744

SN - 364215557X

SN - 9783642155574

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

SP - 748

EP - 761

BT - Computer Vision, ECCV 2010 - 11th European Conference on Computer Vision, Proceedings

PB - Springer Verlag

T2 - 11th European Conference on Computer Vision, ECCV 2010

Y2 - 10 September 2010 through 11 September 2010

ER -