Quantization from Bayes factors with application to multilevel thresholding

F. Murtagh, J. L. Starck

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

We are concerned with the optimal selection of multiple thresholds in image analysis. We propose the use of the Bayes information criterion, a minimal information measure, for this and illustrate its use in practical cases. Applications of multiple threshold selection of interest to us include the closely related problems of (i) quantization for lossy encoding, and (ii) segmentation. Our examples relate to segmentation as a post-processing phase in edge detection.

LanguageEnglish
Pages2001-2007
Number of pages7
JournalPattern Recognition Letters
Volume24
Issue number12
Early online date25 Mar 2003
DOIs
Publication statusPublished - 1 Aug 2003
Externally publishedYes

Fingerprint

Edge detection
Image analysis
Processing

Cite this

@article{fd008b2b7ae64316a80907baab7eb9f1,
title = "Quantization from Bayes factors with application to multilevel thresholding",
abstract = "We are concerned with the optimal selection of multiple thresholds in image analysis. We propose the use of the Bayes information criterion, a minimal information measure, for this and illustrate its use in practical cases. Applications of multiple threshold selection of interest to us include the closely related problems of (i) quantization for lossy encoding, and (ii) segmentation. Our examples relate to segmentation as a post-processing phase in edge detection.",
keywords = "Bayes factor, Bayes information criterion, Edge detection, Image thresholding, Model selection, Wavelet transform",
author = "F. Murtagh and Starck, {J. L.}",
year = "2003",
month = "8",
day = "1",
doi = "10.1016/S0167-8655(03)00038-2",
language = "English",
volume = "24",
pages = "2001--2007",
journal = "Pattern Recognition Letters",
issn = "0167-8655",
publisher = "North-Holland Publ Co",
number = "12",

}

Quantization from Bayes factors with application to multilevel thresholding. / Murtagh, F.; Starck, J. L.

In: Pattern Recognition Letters, Vol. 24, No. 12, 01.08.2003, p. 2001-2007.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Quantization from Bayes factors with application to multilevel thresholding

AU - Murtagh, F.

AU - Starck, J. L.

PY - 2003/8/1

Y1 - 2003/8/1

N2 - We are concerned with the optimal selection of multiple thresholds in image analysis. We propose the use of the Bayes information criterion, a minimal information measure, for this and illustrate its use in practical cases. Applications of multiple threshold selection of interest to us include the closely related problems of (i) quantization for lossy encoding, and (ii) segmentation. Our examples relate to segmentation as a post-processing phase in edge detection.

AB - We are concerned with the optimal selection of multiple thresholds in image analysis. We propose the use of the Bayes information criterion, a minimal information measure, for this and illustrate its use in practical cases. Applications of multiple threshold selection of interest to us include the closely related problems of (i) quantization for lossy encoding, and (ii) segmentation. Our examples relate to segmentation as a post-processing phase in edge detection.

KW - Bayes factor

KW - Bayes information criterion

KW - Edge detection

KW - Image thresholding

KW - Model selection

KW - Wavelet transform

UR - http://www.scopus.com/inward/record.url?scp=0038731223&partnerID=8YFLogxK

U2 - 10.1016/S0167-8655(03)00038-2

DO - 10.1016/S0167-8655(03)00038-2

M3 - Article

VL - 24

SP - 2001

EP - 2007

JO - Pattern Recognition Letters

T2 - Pattern Recognition Letters

JF - Pattern Recognition Letters

SN - 0167-8655

IS - 12

ER -