Framework for texture classification and retrieval using scale invariant feature transform

Tuan Do, Antti Aikala, Olli Saarela

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

2 Citations (Scopus)

Abstract

Texture images can be characterized with key features extracted from images. In this paper, the scale invariant feature transform (hereinafter SIFT) algorithm is utilized to generate local features for texture image classification. The local features are selected as inputs for texture classification framework. For each texture category, a texton dictionary is built based on the local features. To establish the texton dictionary, an adaptive mean shift clustering algorithm is run with all local features to generate key features (called textons) for texton dictionary. The texton dictionaries among texture categories are supposed be distinctive from each other to provide a highest performance in term of classification accuracy. A framework is proposed for classifying images into corresponding categories by matching their local features with textons from the texton dictionaries. This can be done by a histogram model of 'match' vectors versus texture categories. Finally, our texture image database and the Ponce texture database are used to test the proposed approach. The results indicate a potential of our proposed method based on high classification accuracies achieved. They are 100% with our testing database for both classification and retrieval and 92 % and 100% with Ponce database for classification and retrieval, respectively
Original languageEnglish
Title of host publicationProceedings
Subtitle of host publicationNinth International Joint Conference on Computer Science and Software Engineering 2012, JCSSE 2012
PublisherIEEE Institute of Electrical and Electronic Engineers
Pages289-293
ISBN (Electronic)978-1-4673-1921-8
ISBN (Print)978-1-4673-1920-1
DOIs
Publication statusPublished - 2012
MoE publication typeNot Eligible
EventNinth International Joint Conference on Computer Science and Software Engineering 2012, JCSSE'12 - Bangkok, Thailand
Duration: 30 May 20121 Jun 2012

Conference

ConferenceNinth International Joint Conference on Computer Science and Software Engineering 2012, JCSSE'12
Abbreviated titleJCSSE 2012
CountryThailand
CityBangkok
Period30/05/121/06/12

Fingerprint

Textures
Glossaries
Image classification
Clustering algorithms
Testing

Keywords

  • SIFT
  • local feature
  • adaptive mean shift clustering
  • texton
  • texton dictionary

Cite this

Do, T., Aikala, A., & Saarela, O. (2012). Framework for texture classification and retrieval using scale invariant feature transform. In Proceedings: Ninth International Joint Conference on Computer Science and Software Engineering 2012, JCSSE 2012 (pp. 289-293). IEEE Institute of Electrical and Electronic Engineers . https://doi.org/10.1109/JCSSE.2012.6261967
Do, Tuan ; Aikala, Antti ; Saarela, Olli. / Framework for texture classification and retrieval using scale invariant feature transform. Proceedings: Ninth International Joint Conference on Computer Science and Software Engineering 2012, JCSSE 2012. IEEE Institute of Electrical and Electronic Engineers , 2012. pp. 289-293
@inproceedings{b5456bad4b8f4c0d9da8bb07db0ecf78,
title = "Framework for texture classification and retrieval using scale invariant feature transform",
abstract = "Texture images can be characterized with key features extracted from images. In this paper, the scale invariant feature transform (hereinafter SIFT) algorithm is utilized to generate local features for texture image classification. The local features are selected as inputs for texture classification framework. For each texture category, a texton dictionary is built based on the local features. To establish the texton dictionary, an adaptive mean shift clustering algorithm is run with all local features to generate key features (called textons) for texton dictionary. The texton dictionaries among texture categories are supposed be distinctive from each other to provide a highest performance in term of classification accuracy. A framework is proposed for classifying images into corresponding categories by matching their local features with textons from the texton dictionaries. This can be done by a histogram model of 'match' vectors versus texture categories. Finally, our texture image database and the Ponce texture database are used to test the proposed approach. The results indicate a potential of our proposed method based on high classification accuracies achieved. They are 100{\%} with our testing database for both classification and retrieval and 92 {\%} and 100{\%} with Ponce database for classification and retrieval, respectively",
keywords = "SIFT, local feature, adaptive mean shift clustering, texton, texton dictionary",
author = "Tuan Do and Antti Aikala and Olli Saarela",
note = "Project code: 75978 ERCIM Decision Support",
year = "2012",
doi = "10.1109/JCSSE.2012.6261967",
language = "English",
isbn = "978-1-4673-1920-1",
pages = "289--293",
booktitle = "Proceedings",
publisher = "IEEE Institute of Electrical and Electronic Engineers",
address = "United States",

}

Do, T, Aikala, A & Saarela, O 2012, Framework for texture classification and retrieval using scale invariant feature transform. in Proceedings: Ninth International Joint Conference on Computer Science and Software Engineering 2012, JCSSE 2012. IEEE Institute of Electrical and Electronic Engineers , pp. 289-293, Ninth International Joint Conference on Computer Science and Software Engineering 2012, JCSSE'12, Bangkok, Thailand, 30/05/12. https://doi.org/10.1109/JCSSE.2012.6261967

Framework for texture classification and retrieval using scale invariant feature transform. / Do, Tuan; Aikala, Antti; Saarela, Olli.

Proceedings: Ninth International Joint Conference on Computer Science and Software Engineering 2012, JCSSE 2012. IEEE Institute of Electrical and Electronic Engineers , 2012. p. 289-293.

Research output: Chapter in Book/Report/Conference proceedingConference article in proceedingsScientificpeer-review

TY - GEN

T1 - Framework for texture classification and retrieval using scale invariant feature transform

AU - Do, Tuan

AU - Aikala, Antti

AU - Saarela, Olli

N1 - Project code: 75978 ERCIM Decision Support

PY - 2012

Y1 - 2012

N2 - Texture images can be characterized with key features extracted from images. In this paper, the scale invariant feature transform (hereinafter SIFT) algorithm is utilized to generate local features for texture image classification. The local features are selected as inputs for texture classification framework. For each texture category, a texton dictionary is built based on the local features. To establish the texton dictionary, an adaptive mean shift clustering algorithm is run with all local features to generate key features (called textons) for texton dictionary. The texton dictionaries among texture categories are supposed be distinctive from each other to provide a highest performance in term of classification accuracy. A framework is proposed for classifying images into corresponding categories by matching their local features with textons from the texton dictionaries. This can be done by a histogram model of 'match' vectors versus texture categories. Finally, our texture image database and the Ponce texture database are used to test the proposed approach. The results indicate a potential of our proposed method based on high classification accuracies achieved. They are 100% with our testing database for both classification and retrieval and 92 % and 100% with Ponce database for classification and retrieval, respectively

AB - Texture images can be characterized with key features extracted from images. In this paper, the scale invariant feature transform (hereinafter SIFT) algorithm is utilized to generate local features for texture image classification. The local features are selected as inputs for texture classification framework. For each texture category, a texton dictionary is built based on the local features. To establish the texton dictionary, an adaptive mean shift clustering algorithm is run with all local features to generate key features (called textons) for texton dictionary. The texton dictionaries among texture categories are supposed be distinctive from each other to provide a highest performance in term of classification accuracy. A framework is proposed for classifying images into corresponding categories by matching their local features with textons from the texton dictionaries. This can be done by a histogram model of 'match' vectors versus texture categories. Finally, our texture image database and the Ponce texture database are used to test the proposed approach. The results indicate a potential of our proposed method based on high classification accuracies achieved. They are 100% with our testing database for both classification and retrieval and 92 % and 100% with Ponce database for classification and retrieval, respectively

KW - SIFT

KW - local feature

KW - adaptive mean shift clustering

KW - texton

KW - texton dictionary

U2 - 10.1109/JCSSE.2012.6261967

DO - 10.1109/JCSSE.2012.6261967

M3 - Conference article in proceedings

SN - 978-1-4673-1920-1

SP - 289

EP - 293

BT - Proceedings

PB - IEEE Institute of Electrical and Electronic Engineers

ER -

Do T, Aikala A, Saarela O. Framework for texture classification and retrieval using scale invariant feature transform. In Proceedings: Ninth International Joint Conference on Computer Science and Software Engineering 2012, JCSSE 2012. IEEE Institute of Electrical and Electronic Engineers . 2012. p. 289-293 https://doi.org/10.1109/JCSSE.2012.6261967