the Mean Field Theory in EM Procedures for Markov Random Fields

Zhang, Jun; Modestino, J.W.

doi:10.1109/isit.1991.695223

Cited by 55 publications

(80 citation statements)

References 3 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Smoothness of the final segmentation is enforced with a global and stationary Markov Random Field (MRF), which is integrated using the mean field approximation (Zhang, 1992), following the example of Van Leemput et al (1999) and Cardoso et al (2011).…”

Section: Intensity-based Label Refinement Through Expectationmaximisamentioning

confidence: 99%

Robust whole-brain segmentation: Application to traumatic brain injury

Ledig

Heckemann

Hammers

et al. 2015

Medical Image Analysis

160

155

View full text Add to dashboard Cite

We propose a framework for the robust and fully-automatic segmentation of magnetic resonance (MR) brain images called "Multi-Atlas Label Propagation with Expectation-Maximisation based refinement" (MALP-EM). The presented approach is based on a robust registration approach (MAPER), highly performant label fusion (joint label fusion) and intensity-based label refinement using EM. We further adapt this framework to be applicable for the segmentation of brain images with gross changes in anatomy. We propose to account for consistent registration errors by relaxing anatomical priors obtained by multi-atlas propagation and a weighting scheme to locally combine anatomical atlas priors and intensity-refined posterior probabilities. The method is evaluated on a benchmark dataset used in a recent MICCAI segmentation challenge. In this context we show that MALP-EM is competitive for the segmentation of MR brain scans of healthy adults when compared to state-of-the-art automatic labelling techniques. To demonstrate the versatility of the proposed approach, we employed MALP-EM to segment 125 MR brain images into 134 regions from subjects who had sustained traumatic brain injury (TBI). We employ a protocol to assess segmentation quality if no manual reference labels are available. Based on this protocol, three independent, blinded raters confirmed on 13 MR brain scans with pathology that MALP-EM is superior to established label fusion techniques. We visually confirm the robustness of our segmentation approach on the full cohort and investigate the potential of derived symmetry-based imaging biomarkers that correlate with and predict clinically relevant variables in TBI such as the Marshall Classification (MC) or Glasgow Outcome Score (GOS). Specifically, we show that we are able to stratify TBI patients with favourable outcomes from non-favourable outcomes with 64.7% accuracy using acute-phase MR images and 66.8% accuracy using follow-up MR images. Furthermore, we are able to differentiate subjects with the presence of a mass lesion or midline shift from those with diffuse brain injury with 76.0% accuracy. The thalamus, putamen, pallidum and hippocampus are particularly affected. Their involvement predicts TBI disease progression.

show abstract

Section: Intensity-based Label Refinement Through Expectationmaximisamentioning

confidence: 99%

Robust whole-brain segmentation: Application to traumatic brain injury

Ledig

Heckemann

Hammers

et al. 2015

Medical Image Analysis

160

155

View full text Add to dashboard Cite

show abstract

“…In future work we plan to replace the stochastic sampling algorithm by a deterministic scheme as Mean Field method [16] or Belief Propagation [23] method in order to meet the required time constraints.…”

Section: Discussionmentioning

confidence: 99%

“…This is the case here since the partition function cannot be evaluated even for very small size images. One line of research consists in approximating the model in order to obtain a formula where the partition function no longer appears: pseudo-likelihood [14,15], mean field methods [16,17], as well as Bethe trees models [12] are among them. Another possibility is to use stochastic gradient as in [18].…”

Section: Parameter Estimationmentioning

confidence: 99%

Skin detection using pairwise models

Jedynak

Zheng

Daoudi

2005

Image and Vision Computing

View full text Add to dashboard Cite

“…Many combinatorial optimization techniques have been proposed, including iterated conditional mode (ICM) [16], simulated annealing (SA) [13], mean field theory [17], genetic algorithm [18], belief propagation [19], and graph theoretic techniques [20]. To simplify the complexity of the problem, the MRF can be defined on irregular graphs rather than the regular image lattice.…”

Section: Extending To Region-based Segmentationmentioning

confidence: 99%

Preserving objects in Markov Random Fields region growing image segmentation

Dawoud

Netchaev

2011

Pattern Anal Applic

View full text Add to dashboard Cite

This paper proposes an algorithm that preserves objects in Markov Random Fields (MRF) region growing based image segmentation. This is achieved by modifying the MRF energy minimization process so that it would penalize merging regions that have real edges in the boundary between them. Experimental results show that the integration of edge information increases the precision of the segmentation by ensuring the conservation of the objects contours during the region-growing process.

show abstract

the Mean Field Theory in EM Procedures for Markov Random Fields

Abstract: In this paper, we describe how the mean field theory from statistical mechanics can be used in EM (expectation-maximization) algorithm when part of the data is modeled as a Markov random field (MRF).

Cited by 55 publications

References 3 publications

Robust whole-brain segmentation: Application to traumatic brain injury

Robust whole-brain segmentation: Application to traumatic brain injury

Skin detection using pairwise models

Preserving objects in Markov Random Fields region growing image segmentation

Contact Info

Product

Resources

About