LHI-tree: An efficient disk-based image search application

László Havasi, Domonkos Varga, T. Szirányi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

Querying of nearest neighbour (NN) elements on large data collections is an important task for several information or content retrieval tasks. In the paper Local Hash-indexing tree (LHI-tree) is introduced, which is a disk-based index scheme that uses RAM for quick space partition localization and hard disks for the hash indexing. When large collections are considered, such hybrid data structure should be used to implement an effective indexing service. The proposed structure can produce list of approximate nearest neighbours. We compare LHI-tree to FLANN (Fast Library for Approximate Nearest Neighbors), an effective and frequently instanced implementation of ANN search. We show that they produce similar lists of retrieved images (although FLANN works only on RAM). In case of huge multimedia database a disk based indexing and retrieval method has a significant advantage against a vector based system running in RAM data. For the visual content indexing we built an image descriptor composed of four different information representations: edge histogram, entropy histogram, pattern histogram and dominant component colour characteristics. The paper will mention the content based retrieval of Hungarian Wikipedia images.

Original languageEnglish
Title of host publication2014 International Workshop on Computational Intelligence for Multimedia Understanding, IWCIM 2014
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Print)9781479979714
DOIs
Publication statusPublished - Jan 13 2014
Event2014 International Workshop on Computational Intelligence for Multimedia Understanding, IWCIM 2014 - Paris, France
Duration: Nov 1 2014Nov 2 2014

Other

Other2014 International Workshop on Computational Intelligence for Multimedia Understanding, IWCIM 2014
CountryFrance
CityParis
Period11/1/1411/2/14

Fingerprint

Random access storage
Content based retrieval
Hard disk storage
Data structures
Entropy
Color
Nearest neighbor search
Indexing
Nearest Neighbor

Keywords

  • Content based retrieval
  • Disk-based tree
  • Hierarchical tree
  • High dimensional data searching
  • Image descriptor
  • Multimedia indexing

ASJC Scopus subject areas

  • Artificial Intelligence
  • Classics

Cite this

Havasi, L., Varga, D., & Szirányi, T. (2014). LHI-tree: An efficient disk-based image search application. In 2014 International Workshop on Computational Intelligence for Multimedia Understanding, IWCIM 2014 [7008794] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IWCIM.2014.7008794

LHI-tree : An efficient disk-based image search application. / Havasi, László; Varga, Domonkos; Szirányi, T.

2014 International Workshop on Computational Intelligence for Multimedia Understanding, IWCIM 2014. Institute of Electrical and Electronics Engineers Inc., 2014. 7008794.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Havasi, L, Varga, D & Szirányi, T 2014, LHI-tree: An efficient disk-based image search application. in 2014 International Workshop on Computational Intelligence for Multimedia Understanding, IWCIM 2014., 7008794, Institute of Electrical and Electronics Engineers Inc., 2014 International Workshop on Computational Intelligence for Multimedia Understanding, IWCIM 2014, Paris, France, 11/1/14. https://doi.org/10.1109/IWCIM.2014.7008794
Havasi L, Varga D, Szirányi T. LHI-tree: An efficient disk-based image search application. In 2014 International Workshop on Computational Intelligence for Multimedia Understanding, IWCIM 2014. Institute of Electrical and Electronics Engineers Inc. 2014. 7008794 https://doi.org/10.1109/IWCIM.2014.7008794
Havasi, László ; Varga, Domonkos ; Szirányi, T. / LHI-tree : An efficient disk-based image search application. 2014 International Workshop on Computational Intelligence for Multimedia Understanding, IWCIM 2014. Institute of Electrical and Electronics Engineers Inc., 2014.
@inproceedings{3d41943ea2274e0f85a5c972ec671659,
title = "LHI-tree: An efficient disk-based image search application",
abstract = "Querying of nearest neighbour (NN) elements on large data collections is an important task for several information or content retrieval tasks. In the paper Local Hash-indexing tree (LHI-tree) is introduced, which is a disk-based index scheme that uses RAM for quick space partition localization and hard disks for the hash indexing. When large collections are considered, such hybrid data structure should be used to implement an effective indexing service. The proposed structure can produce list of approximate nearest neighbours. We compare LHI-tree to FLANN (Fast Library for Approximate Nearest Neighbors), an effective and frequently instanced implementation of ANN search. We show that they produce similar lists of retrieved images (although FLANN works only on RAM). In case of huge multimedia database a disk based indexing and retrieval method has a significant advantage against a vector based system running in RAM data. For the visual content indexing we built an image descriptor composed of four different information representations: edge histogram, entropy histogram, pattern histogram and dominant component colour characteristics. The paper will mention the content based retrieval of Hungarian Wikipedia images.",
keywords = "Content based retrieval, Disk-based tree, Hierarchical tree, High dimensional data searching, Image descriptor, Multimedia indexing",
author = "L{\'a}szl{\'o} Havasi and Domonkos Varga and T. Szir{\'a}nyi",
year = "2014",
month = "1",
day = "13",
doi = "10.1109/IWCIM.2014.7008794",
language = "English",
isbn = "9781479979714",
booktitle = "2014 International Workshop on Computational Intelligence for Multimedia Understanding, IWCIM 2014",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - LHI-tree

T2 - An efficient disk-based image search application

AU - Havasi, László

AU - Varga, Domonkos

AU - Szirányi, T.

PY - 2014/1/13

Y1 - 2014/1/13

N2 - Querying of nearest neighbour (NN) elements on large data collections is an important task for several information or content retrieval tasks. In the paper Local Hash-indexing tree (LHI-tree) is introduced, which is a disk-based index scheme that uses RAM for quick space partition localization and hard disks for the hash indexing. When large collections are considered, such hybrid data structure should be used to implement an effective indexing service. The proposed structure can produce list of approximate nearest neighbours. We compare LHI-tree to FLANN (Fast Library for Approximate Nearest Neighbors), an effective and frequently instanced implementation of ANN search. We show that they produce similar lists of retrieved images (although FLANN works only on RAM). In case of huge multimedia database a disk based indexing and retrieval method has a significant advantage against a vector based system running in RAM data. For the visual content indexing we built an image descriptor composed of four different information representations: edge histogram, entropy histogram, pattern histogram and dominant component colour characteristics. The paper will mention the content based retrieval of Hungarian Wikipedia images.

AB - Querying of nearest neighbour (NN) elements on large data collections is an important task for several information or content retrieval tasks. In the paper Local Hash-indexing tree (LHI-tree) is introduced, which is a disk-based index scheme that uses RAM for quick space partition localization and hard disks for the hash indexing. When large collections are considered, such hybrid data structure should be used to implement an effective indexing service. The proposed structure can produce list of approximate nearest neighbours. We compare LHI-tree to FLANN (Fast Library for Approximate Nearest Neighbors), an effective and frequently instanced implementation of ANN search. We show that they produce similar lists of retrieved images (although FLANN works only on RAM). In case of huge multimedia database a disk based indexing and retrieval method has a significant advantage against a vector based system running in RAM data. For the visual content indexing we built an image descriptor composed of four different information representations: edge histogram, entropy histogram, pattern histogram and dominant component colour characteristics. The paper will mention the content based retrieval of Hungarian Wikipedia images.

KW - Content based retrieval

KW - Disk-based tree

KW - Hierarchical tree

KW - High dimensional data searching

KW - Image descriptor

KW - Multimedia indexing

UR - http://www.scopus.com/inward/record.url?scp=84946686379&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84946686379&partnerID=8YFLogxK

U2 - 10.1109/IWCIM.2014.7008794

DO - 10.1109/IWCIM.2014.7008794

M3 - Conference contribution

AN - SCOPUS:84946686379

SN - 9781479979714

BT - 2014 International Workshop on Computational Intelligence for Multimedia Understanding, IWCIM 2014

PB - Institute of Electrical and Electronics Engineers Inc.

ER -