Fixed partitioning and salient points with MPEG-7 cluster correlograms for image categorization

Azizi Abdullah, Remco C. Veltkamp, Marco A. Wiering

Research output: Contribution to journalArticle

24 Citations (Scopus)

Abstract

This paper compares fixed partitioning and salient points schemes for dividing an image into patches, in combination with low-level MPEG-7 visual descriptors to represent the patches with particular patterns. A clustering technique is applied to construct a compact representation by grouping similar patterns into a cluster codebook. The codebook will then be used to encode the patterns into visual keywords. In order to obtain high-level information about the relational context of an image, a correlogram is constructed from the spatial relations between visual keyword indices in an image. For classifying images a k-nearest neighbors (k-NN) and a support vector machine (SVM) algorithm are used and compared. The techniques are compared to other methods on two well-known datasets, namely Corel and PASCAL. To measure the performance of the proposed algorithms, average precision, a confusion matrix, and ROC-curves are used. The results show that the cluster correlogram outperforms the cluster histogram. The saliency based scheme performs similarly to the fixed partitioning scheme and the SVM significantly outperforms the k-NN classifier. Finally, we demonstrate the robustness to noise, photometric, and geometric distortions.

Original languageEnglish
Pages (from-to)650-662
Number of pages13
JournalPattern Recognition
Volume43
Issue number3
DOIs
Publication statusPublished - Mar 2010
Externally publishedYes

Fingerprint

Support vector machines
Classifiers

Keywords

  • Cluster correlogram
  • Computer vision
  • Image indexing and retrieval

ASJC Scopus subject areas

  • Software
  • Artificial Intelligence
  • Computer Vision and Pattern Recognition
  • Signal Processing

Cite this

Fixed partitioning and salient points with MPEG-7 cluster correlograms for image categorization. / Abdullah, Azizi; Veltkamp, Remco C.; Wiering, Marco A.

In: Pattern Recognition, Vol. 43, No. 3, 03.2010, p. 650-662.

Research output: Contribution to journalArticle

Abdullah, Azizi ; Veltkamp, Remco C. ; Wiering, Marco A. / Fixed partitioning and salient points with MPEG-7 cluster correlograms for image categorization. In: Pattern Recognition. 2010 ; Vol. 43, No. 3. pp. 650-662.
@article{5d5353242f0e465aa65bf40463fdc5c0,
title = "Fixed partitioning and salient points with MPEG-7 cluster correlograms for image categorization",
abstract = "This paper compares fixed partitioning and salient points schemes for dividing an image into patches, in combination with low-level MPEG-7 visual descriptors to represent the patches with particular patterns. A clustering technique is applied to construct a compact representation by grouping similar patterns into a cluster codebook. The codebook will then be used to encode the patterns into visual keywords. In order to obtain high-level information about the relational context of an image, a correlogram is constructed from the spatial relations between visual keyword indices in an image. For classifying images a k-nearest neighbors (k-NN) and a support vector machine (SVM) algorithm are used and compared. The techniques are compared to other methods on two well-known datasets, namely Corel and PASCAL. To measure the performance of the proposed algorithms, average precision, a confusion matrix, and ROC-curves are used. The results show that the cluster correlogram outperforms the cluster histogram. The saliency based scheme performs similarly to the fixed partitioning scheme and the SVM significantly outperforms the k-NN classifier. Finally, we demonstrate the robustness to noise, photometric, and geometric distortions.",
keywords = "Cluster correlogram, Computer vision, Image indexing and retrieval",
author = "Azizi Abdullah and Veltkamp, {Remco C.} and Wiering, {Marco A.}",
year = "2010",
month = "3",
doi = "10.1016/j.patcog.2009.09.007",
language = "English",
volume = "43",
pages = "650--662",
journal = "Pattern Recognition",
issn = "0031-3203",
publisher = "Elsevier Limited",
number = "3",

}

TY - JOUR

T1 - Fixed partitioning and salient points with MPEG-7 cluster correlograms for image categorization

AU - Abdullah, Azizi

AU - Veltkamp, Remco C.

AU - Wiering, Marco A.

PY - 2010/3

Y1 - 2010/3

N2 - This paper compares fixed partitioning and salient points schemes for dividing an image into patches, in combination with low-level MPEG-7 visual descriptors to represent the patches with particular patterns. A clustering technique is applied to construct a compact representation by grouping similar patterns into a cluster codebook. The codebook will then be used to encode the patterns into visual keywords. In order to obtain high-level information about the relational context of an image, a correlogram is constructed from the spatial relations between visual keyword indices in an image. For classifying images a k-nearest neighbors (k-NN) and a support vector machine (SVM) algorithm are used and compared. The techniques are compared to other methods on two well-known datasets, namely Corel and PASCAL. To measure the performance of the proposed algorithms, average precision, a confusion matrix, and ROC-curves are used. The results show that the cluster correlogram outperforms the cluster histogram. The saliency based scheme performs similarly to the fixed partitioning scheme and the SVM significantly outperforms the k-NN classifier. Finally, we demonstrate the robustness to noise, photometric, and geometric distortions.

AB - This paper compares fixed partitioning and salient points schemes for dividing an image into patches, in combination with low-level MPEG-7 visual descriptors to represent the patches with particular patterns. A clustering technique is applied to construct a compact representation by grouping similar patterns into a cluster codebook. The codebook will then be used to encode the patterns into visual keywords. In order to obtain high-level information about the relational context of an image, a correlogram is constructed from the spatial relations between visual keyword indices in an image. For classifying images a k-nearest neighbors (k-NN) and a support vector machine (SVM) algorithm are used and compared. The techniques are compared to other methods on two well-known datasets, namely Corel and PASCAL. To measure the performance of the proposed algorithms, average precision, a confusion matrix, and ROC-curves are used. The results show that the cluster correlogram outperforms the cluster histogram. The saliency based scheme performs similarly to the fixed partitioning scheme and the SVM significantly outperforms the k-NN classifier. Finally, we demonstrate the robustness to noise, photometric, and geometric distortions.

KW - Cluster correlogram

KW - Computer vision

KW - Image indexing and retrieval

UR - http://www.scopus.com/inward/record.url?scp=70449699633&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70449699633&partnerID=8YFLogxK

U2 - 10.1016/j.patcog.2009.09.007

DO - 10.1016/j.patcog.2009.09.007

M3 - Article

VL - 43

SP - 650

EP - 662

JO - Pattern Recognition

JF - Pattern Recognition

SN - 0031-3203

IS - 3

ER -