Most edges in Markov random fields for white matter hyperintensity segmentation are worthless

Christopher Schwarz, Evan Fletcher, Baljeet Singh, Amy Liu, Noel Smith, Charles Decarli, Owen Carmichael

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

Abstract

The time and space complexities of Markov random field (MRF) algorithms for image segmentation increase with the number of edges that represent statistical dependencies between adjacent pixels. This has made MRFs too computationally complex for cutting-edge applications such as joint segmentation of longitudinal sequences of many high-resolution magnetic resonance images (MRIs). Here, we show that simply removing edges from full MRFs can reduce the computational complexity of MRF parameter estimation and inference with no notable decrease in segmentation performance. In particular, we show that for segmentation of white matter hyperintensities in 88 brain MRI scans of elderly individuals, as many as 66% of MRF edges can be removed without substantially degrading segmentation accuracy. We then show that removing edges from MRFs makes MRF parameter estimation and inference computationally tractable enough to enable modeling statistical dependencies within and across a larger number of brain MRI scans in a longitudinal series; this improves segmentation performance compared to separate segmentations of each individual scan in the series.

Original languageEnglish (US)
Title of host publication2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2012
Pages2684-2687
Number of pages4
DOIs
StatePublished - Dec 14 2012
Externally publishedYes
Event34th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS 2012 - San Diego, CA, United States
Duration: Aug 28 2012Sep 1 2012

Publication series

NameProceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS
ISSN (Print)1557-170X

Other

Other34th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS 2012
CountryUnited States
CitySan Diego, CA
Period8/28/129/1/12

Fingerprint

Magnetic resonance
Magnetic Resonance Spectroscopy
Parameter estimation
Brain
Image segmentation
Computational complexity
Joints
Pixels
White Matter

ASJC Scopus subject areas

  • Signal Processing
  • Biomedical Engineering
  • Computer Vision and Pattern Recognition
  • Health Informatics

Cite this

Schwarz, C., Fletcher, E., Singh, B., Liu, A., Smith, N., Decarli, C., & Carmichael, O. (2012). Most edges in Markov random fields for white matter hyperintensity segmentation are worthless. In 2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2012 (pp. 2684-2687). [6346517] (Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS). https://doi.org/10.1109/EMBC.2012.6346517

Most edges in Markov random fields for white matter hyperintensity segmentation are worthless. / Schwarz, Christopher; Fletcher, Evan; Singh, Baljeet; Liu, Amy; Smith, Noel; Decarli, Charles; Carmichael, Owen.

2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2012. 2012. p. 2684-2687 6346517 (Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS).

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

Schwarz, C, Fletcher, E, Singh, B, Liu, A, Smith, N, Decarli, C & Carmichael, O 2012, Most edges in Markov random fields for white matter hyperintensity segmentation are worthless. in 2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2012., 6346517, Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS, pp. 2684-2687, 34th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS 2012, San Diego, CA, United States, 8/28/12. https://doi.org/10.1109/EMBC.2012.6346517
Schwarz C, Fletcher E, Singh B, Liu A, Smith N, Decarli C et al. Most edges in Markov random fields for white matter hyperintensity segmentation are worthless. In 2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2012. 2012. p. 2684-2687. 6346517. (Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS). https://doi.org/10.1109/EMBC.2012.6346517
Schwarz, Christopher ; Fletcher, Evan ; Singh, Baljeet ; Liu, Amy ; Smith, Noel ; Decarli, Charles ; Carmichael, Owen. / Most edges in Markov random fields for white matter hyperintensity segmentation are worthless. 2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2012. 2012. pp. 2684-2687 (Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS).
@inproceedings{7b1d8b1237c04aa6a78628f3dc3c9b5f,
title = "Most edges in Markov random fields for white matter hyperintensity segmentation are worthless",
abstract = "The time and space complexities of Markov random field (MRF) algorithms for image segmentation increase with the number of edges that represent statistical dependencies between adjacent pixels. This has made MRFs too computationally complex for cutting-edge applications such as joint segmentation of longitudinal sequences of many high-resolution magnetic resonance images (MRIs). Here, we show that simply removing edges from full MRFs can reduce the computational complexity of MRF parameter estimation and inference with no notable decrease in segmentation performance. In particular, we show that for segmentation of white matter hyperintensities in 88 brain MRI scans of elderly individuals, as many as 66{\%} of MRF edges can be removed without substantially degrading segmentation accuracy. We then show that removing edges from MRFs makes MRF parameter estimation and inference computationally tractable enough to enable modeling statistical dependencies within and across a larger number of brain MRI scans in a longitudinal series; this improves segmentation performance compared to separate segmentations of each individual scan in the series.",
author = "Christopher Schwarz and Evan Fletcher and Baljeet Singh and Amy Liu and Noel Smith and Charles Decarli and Owen Carmichael",
year = "2012",
month = "12",
day = "14",
doi = "10.1109/EMBC.2012.6346517",
language = "English (US)",
isbn = "9781424441198",
series = "Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS",
pages = "2684--2687",
booktitle = "2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2012",

}

TY - GEN

T1 - Most edges in Markov random fields for white matter hyperintensity segmentation are worthless

AU - Schwarz, Christopher

AU - Fletcher, Evan

AU - Singh, Baljeet

AU - Liu, Amy

AU - Smith, Noel

AU - Decarli, Charles

AU - Carmichael, Owen

PY - 2012/12/14

Y1 - 2012/12/14

N2 - The time and space complexities of Markov random field (MRF) algorithms for image segmentation increase with the number of edges that represent statistical dependencies between adjacent pixels. This has made MRFs too computationally complex for cutting-edge applications such as joint segmentation of longitudinal sequences of many high-resolution magnetic resonance images (MRIs). Here, we show that simply removing edges from full MRFs can reduce the computational complexity of MRF parameter estimation and inference with no notable decrease in segmentation performance. In particular, we show that for segmentation of white matter hyperintensities in 88 brain MRI scans of elderly individuals, as many as 66% of MRF edges can be removed without substantially degrading segmentation accuracy. We then show that removing edges from MRFs makes MRF parameter estimation and inference computationally tractable enough to enable modeling statistical dependencies within and across a larger number of brain MRI scans in a longitudinal series; this improves segmentation performance compared to separate segmentations of each individual scan in the series.

AB - The time and space complexities of Markov random field (MRF) algorithms for image segmentation increase with the number of edges that represent statistical dependencies between adjacent pixels. This has made MRFs too computationally complex for cutting-edge applications such as joint segmentation of longitudinal sequences of many high-resolution magnetic resonance images (MRIs). Here, we show that simply removing edges from full MRFs can reduce the computational complexity of MRF parameter estimation and inference with no notable decrease in segmentation performance. In particular, we show that for segmentation of white matter hyperintensities in 88 brain MRI scans of elderly individuals, as many as 66% of MRF edges can be removed without substantially degrading segmentation accuracy. We then show that removing edges from MRFs makes MRF parameter estimation and inference computationally tractable enough to enable modeling statistical dependencies within and across a larger number of brain MRI scans in a longitudinal series; this improves segmentation performance compared to separate segmentations of each individual scan in the series.

UR - http://www.scopus.com/inward/record.url?scp=84883007920&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84883007920&partnerID=8YFLogxK

U2 - 10.1109/EMBC.2012.6346517

DO - 10.1109/EMBC.2012.6346517

M3 - Conference contribution

C2 - 23366478

AN - SCOPUS:84883007920

SN - 9781424441198

T3 - Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS

SP - 2684

EP - 2687

BT - 2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2012

ER -