MDPV - Metric Distance Permutation Vocabulary

J 2015

MDPV - Metric Distance Permutation Vocabulary

DOHNAL, Vlastislav, Tomáš HOMOLA and Pavel ZEZULA

Basic information

Original name

MDPV - Metric Distance Permutation Vocabulary

Name in Czech

MDPV - Vizuální slovník založený na permutacích vzdáleností

Authors

DOHNAL, Vlastislav (203 Czech Republic, guarantor, belonging to the institution), Tomáš HOMOLA (203 Czech Republic, belonging to the institution) and Pavel ZEZULA (203 Czech Republic, belonging to the institution)

Edition

Information Retrieval, Netherlands, Springer, 2015, 1386-4564

Other information

Language

English

Type of outcome

Článek v odborném periodiku

Field of Study

10201 Computer sciences, information science, bioinformatics

Country of publisher

Netherlands

Confidentiality degree

není předmětem státního či obchodního tajemství

References:

URL

Impact factor

Impact factor: 0.896

RIV identification code

RIV/00216224:14330/15:00080599

Organization unit

Faculty of Informatics

DOI

http://dx.doi.org/10.1007/s10791-014-9247-6

UT WoS

000348350600003

Keywords in English

feature quantization; visual vocabulary; bag-of-features model; k-means clustering; metric distance permutation vocabulary

Abstract

V originále

Sub-image content-based similarity search forms an important operation in current image archives since it provides users with images that contain a query image as their part. Such a search can conveniently be implemented using the bag-of-features model. Its integral part is a construction of visual vocabulary. Most existing algorithms to create a visual vocabulary suffer from high computational (e.g. k-means) or supervisor-guidance (e.g. visual-bit classifier, or sparse coding) requirements. In this paper, we propose a~novel approach to visual vocabulary construction called Metric Distance Permutation Vocabulary (MDPV). It is based on permutations of metric distances to create compact visual words. Its major advantage over prior techniques is time and space efficiency of vocabulary construction and quantization process during querying, while achieving comparable or even better effectiveness (query result quality). Moreover, this basic concept is extended to combine more independent permutations. Both the proposals are experimented on well-known real-world data-sets and compared to other state-of-the-art techniques.

Links

GBP103/12/G084, research and development project

Name: Centrum pro multi-modální interpretaci dat velkého rozsahu

Investor: Czech Science Foundation

Citovat

DOHNAL, Vlastislav, Tomáš HOMOLA and Pavel ZEZULA. MDPV - Metric Distance Permutation Vocabulary. Information Retrieval. Netherlands: Springer, 2015, vol. 18, No 1, p. 51-72. ISSN 1386-4564. Available from: https://dx.doi.org/10.1007/s10791-014-9247-6.

@article{1201814,
   author = {Dohnal, Vlastislav and Homola, Tomáš and Zezula, Pavel},
   article_location = {Netherlands},
   article_number = {1},
   doi = {http://dx.doi.org/10.1007/s10791-014-9247-6},
   keywords = {feature quantization; visual vocabulary; bag-of-features model; k-means clustering; metric distance permutation vocabulary},
   language = {eng},
   issn = {1386-4564},
   journal = {Information Retrieval},
   title = {MDPV - Metric Distance Permutation Vocabulary},
   url = {http://link.springer.com/article/10.1007/s10791-014-9247-6},
   volume = {18},
   year = {2015}
}

TY  - JOUR
ID  - 1201814
AU  - Dohnal, Vlastislav - Homola, Tomáš - Zezula, Pavel
PY  - 2015
TI  - MDPV - Metric Distance Permutation Vocabulary
JF  - Information Retrieval
VL  - 18
IS  - 1
SP  - 51-72
EP  - 51-72
PB  - Springer
SN  - 13864564
KW  - feature quantization
KW  - visual vocabulary
KW  - bag-of-features model
KW  - k-means clustering
KW  - metric distance permutation vocabulary
UR  - http://link.springer.com/article/10.1007/s10791-014-9247-6
L2  - http://link.springer.com/article/10.1007/s10791-014-9247-6
N2  - Sub-image content-based similarity search forms an important operation in current image archives since it provides users with images that contain a query image as their part. Such a search can conveniently be implemented using the bag-of-features model. Its integral part is a construction of visual vocabulary. Most existing algorithms to create a visual vocabulary suffer from high computational (e.g. k-means) or supervisor-guidance (e.g. visual-bit classifier, or sparse coding) requirements. In this paper, we propose a~novel approach to visual vocabulary construction called Metric Distance Permutation Vocabulary (MDPV). It is based on permutations of metric distances to create compact visual words. Its major advantage over prior techniques is time and space efficiency of vocabulary construction and quantization process during querying, while achieving comparable or even better effectiveness (query result quality). Moreover, this basic concept is extended to combine more independent permutations. Both the proposals are experimented on well-known real-world data-sets and compared to other state-of-the-art techniques.
ER  -

DOHNAL, Vlastislav, Tomáš HOMOLA and Pavel ZEZULA. MDPV - Metric Distance Permutation Vocabulary. \textit{Information Retrieval}. Netherlands: Springer, 2015, vol.~18, No~1, p.~51-72. ISSN~1386-4564. Available from: https://dx.doi.org/10.1007/s10791-014-9247-6.

Detailed Information on Publication Record