Near-duplicate detection using a new framework of constructing accurate affine invariant regions

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

Abstract

In this study, we propose a simple, yet general and powerful framework for constructing accurate affine invariant regions and use it for near-duplicate detection problem. In our framework, a method for extracting reliable seed points is first proposed. Then, regions which are invariant to most common affine transformations are extracted from seed points by a new method named the Thresholding Seeded Growing Region (TSGR). After that, an improved ellipse fitting method based on the Direct Least Square Fitting (DLSF) is used to fit the irregularly-shaped contours of TSGRs to obtain ellipse regions as the final invariant regions. At last, SIFT-PCA descriptors are computed on the obtained regions. In the experiment, our framework is evaluated by retrieving near-duplicate in an image database containing 1000 images. It gives a satisfying result of 96.8% precision at 100% recall.

Original languageEnglish
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages61-72
Number of pages12
Volume4781 LNCS
Publication statusPublished - 2007
Event9th International Conference on Visual Information Systems, VISUAL 2007 - Shanghai
Duration: 2007 Jun 282007 Jun 29

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume4781 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other9th International Conference on Visual Information Systems, VISUAL 2007
CityShanghai
Period07/6/2807/6/29

Fingerprint

Invariant Region
Affine Invariant
Seed
Ellipse
Seeds
Least Square Fitting
Region Growing
Passive Cutaneous Anaphylaxis
Scale Invariant Feature Transform
Image Database
Thresholding
Least-Squares Analysis
Descriptors
Affine transformation
Databases
Invariant
Experiments
Experiment
Framework

Keywords

  • Ellipse fitting
  • Image matching
  • Invariant region
  • Near-duplicate detection
  • Thresholding seeded growing regions

ASJC Scopus subject areas

  • Computer Science(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Theoretical Computer Science

Cite this

Tian, L., & Kamata, S. (2007). Near-duplicate detection using a new framework of constructing accurate affine invariant regions. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 4781 LNCS, pp. 61-72). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4781 LNCS).

Near-duplicate detection using a new framework of constructing accurate affine invariant regions. / Tian, Li; Kamata, Seiichiro.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4781 LNCS 2007. p. 61-72 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4781 LNCS).

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

Tian, L & Kamata, S 2007, Near-duplicate detection using a new framework of constructing accurate affine invariant regions. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). vol. 4781 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 4781 LNCS, pp. 61-72, 9th International Conference on Visual Information Systems, VISUAL 2007, Shanghai, 07/6/28.
Tian L, Kamata S. Near-duplicate detection using a new framework of constructing accurate affine invariant regions. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4781 LNCS. 2007. p. 61-72. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
Tian, Li ; Kamata, Seiichiro. / Near-duplicate detection using a new framework of constructing accurate affine invariant regions. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4781 LNCS 2007. pp. 61-72 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{789abadbcb784e54abfcfc0939989447,
title = "Near-duplicate detection using a new framework of constructing accurate affine invariant regions",
abstract = "In this study, we propose a simple, yet general and powerful framework for constructing accurate affine invariant regions and use it for near-duplicate detection problem. In our framework, a method for extracting reliable seed points is first proposed. Then, regions which are invariant to most common affine transformations are extracted from seed points by a new method named the Thresholding Seeded Growing Region (TSGR). After that, an improved ellipse fitting method based on the Direct Least Square Fitting (DLSF) is used to fit the irregularly-shaped contours of TSGRs to obtain ellipse regions as the final invariant regions. At last, SIFT-PCA descriptors are computed on the obtained regions. In the experiment, our framework is evaluated by retrieving near-duplicate in an image database containing 1000 images. It gives a satisfying result of 96.8{\%} precision at 100{\%} recall.",
keywords = "Ellipse fitting, Image matching, Invariant region, Near-duplicate detection, Thresholding seeded growing regions",
author = "Li Tian and Seiichiro Kamata",
year = "2007",
language = "English",
isbn = "9783540764137",
volume = "4781 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "61--72",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

}

TY - GEN

T1 - Near-duplicate detection using a new framework of constructing accurate affine invariant regions

AU - Tian, Li

AU - Kamata, Seiichiro

PY - 2007

Y1 - 2007

N2 - In this study, we propose a simple, yet general and powerful framework for constructing accurate affine invariant regions and use it for near-duplicate detection problem. In our framework, a method for extracting reliable seed points is first proposed. Then, regions which are invariant to most common affine transformations are extracted from seed points by a new method named the Thresholding Seeded Growing Region (TSGR). After that, an improved ellipse fitting method based on the Direct Least Square Fitting (DLSF) is used to fit the irregularly-shaped contours of TSGRs to obtain ellipse regions as the final invariant regions. At last, SIFT-PCA descriptors are computed on the obtained regions. In the experiment, our framework is evaluated by retrieving near-duplicate in an image database containing 1000 images. It gives a satisfying result of 96.8% precision at 100% recall.

AB - In this study, we propose a simple, yet general and powerful framework for constructing accurate affine invariant regions and use it for near-duplicate detection problem. In our framework, a method for extracting reliable seed points is first proposed. Then, regions which are invariant to most common affine transformations are extracted from seed points by a new method named the Thresholding Seeded Growing Region (TSGR). After that, an improved ellipse fitting method based on the Direct Least Square Fitting (DLSF) is used to fit the irregularly-shaped contours of TSGRs to obtain ellipse regions as the final invariant regions. At last, SIFT-PCA descriptors are computed on the obtained regions. In the experiment, our framework is evaluated by retrieving near-duplicate in an image database containing 1000 images. It gives a satisfying result of 96.8% precision at 100% recall.

KW - Ellipse fitting

KW - Image matching

KW - Invariant region

KW - Near-duplicate detection

KW - Thresholding seeded growing regions

UR - http://www.scopus.com/inward/record.url?scp=38349030529&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=38349030529&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9783540764137

VL - 4781 LNCS

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

SP - 61

EP - 72

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

ER -