Non-parametric estimation of mixture model order

Enrique Corona; Brian Nutter; Sunanda Mitra

doi:10.1109/SSIAI.2008.4512306

Non-parametric estimation of mixture model order

Enrique Corona, Brian Nutter, Sunanda Mitra

Electrical and Computer Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

4 Scopus citations

Abstract

Mixture models are among the most popular and effective techniques for image segmentation. While Gaussian Mixture Models (GMM) are a reasonable choice, the number of components is not easy to determine. A non-parametric technique, based on the transformation and analysis of the D(R) (distortion-rate) curve is proposed for model order identification purposes. This curve is estimated via the popular K-means clustering algorithm. To achieve repeatability and efficiency, various centroid initialization and image down sampling methods are proposed and tested. This technique also provides good starting points for inferring the GMM parameters via the expectation-maximization (EM) algorithm, which effectively reduces the segmentation time and the chances of getting trapped in local optima.

Original language	English
Title of host publication	2008 IEEE Southwest Symposium on Image Analysis and Interpretation, SSIAI 2008 - Proceedings
Pages	145-148
Number of pages	4
DOIs	https://doi.org/10.1109/SSIAI.2008.4512306
State	Published - 2008
Event	2008 IEEE Southwest Symposium on Image Analysis and Interpretation, SSIAI 2008 - Santa Fe, NM, United States Duration: Mar 24 2008 → Mar 26 2008

Publication series

Name	Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation

Conference

Conference	2008 IEEE Southwest Symposium on Image Analysis and Interpretation, SSIAI 2008
Country/Territory	United States
City	Santa Fe, NM
Period	03/24/08 → 03/26/08

Access to Document

10.1109/SSIAI.2008.4512306

Cite this

@inproceedings{19489ed6944f4f0282d2be6998109331,

title = "Non-parametric estimation of mixture model order",

abstract = "Mixture models are among the most popular and effective techniques for image segmentation. While Gaussian Mixture Models (GMM) are a reasonable choice, the number of components is not easy to determine. A non-parametric technique, based on the transformation and analysis of the D(R) (distortion-rate) curve is proposed for model order identification purposes. This curve is estimated via the popular K-means clustering algorithm. To achieve repeatability and efficiency, various centroid initialization and image down sampling methods are proposed and tested. This technique also provides good starting points for inferring the GMM parameters via the expectation-maximization (EM) algorithm, which effectively reduces the segmentation time and the chances of getting trapped in local optima.",

author = "Enrique Corona and Brian Nutter and Sunanda Mitra",

year = "2008",

doi = "10.1109/SSIAI.2008.4512306",

language = "English",

isbn = "9781424422975",

series = "Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation",

pages = "145--148",

booktitle = "2008 IEEE Southwest Symposium on Image Analysis and Interpretation, SSIAI 2008 - Proceedings",

note = "2008 IEEE Southwest Symposium on Image Analysis and Interpretation, SSIAI 2008 ; Conference date: 24-03-2008 Through 26-03-2008",

}

Corona, E, Nutter, B & Mitra, S 2008, Non-parametric estimation of mixture model order. in 2008 IEEE Southwest Symposium on Image Analysis and Interpretation, SSIAI 2008 - Proceedings., 4512306, Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation, pp. 145-148, 2008 IEEE Southwest Symposium on Image Analysis and Interpretation, SSIAI 2008, Santa Fe, NM, United States, 03/24/08. https://doi.org/10.1109/SSIAI.2008.4512306

Non-parametric estimation of mixture model order. / Corona, Enrique; Nutter, Brian; Mitra, Sunanda.
2008 IEEE Southwest Symposium on Image Analysis and Interpretation, SSIAI 2008 - Proceedings. 2008. p. 145-148 4512306 (Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Non-parametric estimation of mixture model order

AU - Corona, Enrique

AU - Nutter, Brian

AU - Mitra, Sunanda

PY - 2008

Y1 - 2008

N2 - Mixture models are among the most popular and effective techniques for image segmentation. While Gaussian Mixture Models (GMM) are a reasonable choice, the number of components is not easy to determine. A non-parametric technique, based on the transformation and analysis of the D(R) (distortion-rate) curve is proposed for model order identification purposes. This curve is estimated via the popular K-means clustering algorithm. To achieve repeatability and efficiency, various centroid initialization and image down sampling methods are proposed and tested. This technique also provides good starting points for inferring the GMM parameters via the expectation-maximization (EM) algorithm, which effectively reduces the segmentation time and the chances of getting trapped in local optima.

AB - Mixture models are among the most popular and effective techniques for image segmentation. While Gaussian Mixture Models (GMM) are a reasonable choice, the number of components is not easy to determine. A non-parametric technique, based on the transformation and analysis of the D(R) (distortion-rate) curve is proposed for model order identification purposes. This curve is estimated via the popular K-means clustering algorithm. To achieve repeatability and efficiency, various centroid initialization and image down sampling methods are proposed and tested. This technique also provides good starting points for inferring the GMM parameters via the expectation-maximization (EM) algorithm, which effectively reduces the segmentation time and the chances of getting trapped in local optima.

UR - http://www.scopus.com/inward/record.url?scp=62349104541&partnerID=8YFLogxK

U2 - 10.1109/SSIAI.2008.4512306

DO - 10.1109/SSIAI.2008.4512306

M3 - Conference contribution

AN - SCOPUS:62349104541

SN - 9781424422975

T3 - Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation

SP - 145

EP - 148

BT - 2008 IEEE Southwest Symposium on Image Analysis and Interpretation, SSIAI 2008 - Proceedings

T2 - 2008 IEEE Southwest Symposium on Image Analysis and Interpretation, SSIAI 2008

Y2 - 24 March 2008 through 26 March 2008

ER -

Non-parametric estimation of mixture model order

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this