Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Variational Bayes inference of spatial mixture models for segmentation.

Mark W Woolrich1, Timothy E Behrens

  • 1Oxford Centre for Functional Magnetic Resonance Imaging of the Brain (FMRIB), University of Oxford, John Radcliffe Hospital, Oxford OX3 9DU, UK. woolrich@fmrib.ox.ac.uk

IEEE Transactions on Medical Imaging
|October 10, 2006
PubMed
Summary

This study introduces a faster inference method for adaptive spatial mixture models used in image segmentation. The new approach significantly speeds up analysis for large datasets, improving medical image segmentation and functional imaging analysis.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Modelling discrete states and long-term dynamics in functional brain networks.

Imaging neuroscience (Cambridge, Mass.)·2026
Same author

Canonical Hidden Markov Model Networks for studying M/EEG.

Imaging neuroscience (Cambridge, Mass.)·2026
Same author

Effects of Age on Resting-State Cortical Networks.

Human brain mapping·2026
Same author

Rethinking how research is reviewed and published.

eLife·2026
Same author

The role of age in the relationship between brain structure and cognition: moderator or confound?

Cerebral cortex (New York, N.Y. : 1991)·2026
Same author

Flexible neural representations of abstract structural knowledge in the human entorhinal cortex.

eLife·2026

Area of Science:

  • Medical image analysis
  • Statistical modeling
  • Computational neuroscience

Background:

  • Mixture models are vital for image segmentation, including medical imaging and functional brain imaging.
  • Spatial mixture models incorporate Markov random fields (MRFs) for regularization.
  • Previous adaptive models relied on slow Markov Chain Monte Carlo (MCMC) sampling for inference.

Purpose of the Study:

  • To develop a more efficient inference method for fully adaptive spatial mixture models.
  • To overcome the computational limitations of MCMC sampling for large-scale image data.

Main Methods:

  • A variational Bayes approximation was combined with a second-order Taylor expansion.
  • This approach addresses intractable components of the posterior distribution in Variational Bayes.

Related Experiment Videos

  • The method enables adaptive parameter determination for spatial regularization.

    • Main Results:

    • Inference speed was improved by an order of magnitude compared to MCMC.
    • The novel method demonstrated efficient inference for adaptive spatial mixture models.
    • Performance was evaluated on artificial data and functional magnetic resonance imaging (fMRI) data.

    Conclusions:

    • The proposed variational Bayes approach offers a computationally efficient alternative to MCMC for adaptive spatial mixture models.
    • This advancement facilitates more rapid and scalable image segmentation in medical and neuroscience applications.
    • The method is suitable for analyzing complex spatial data, including fMRI statistical parametric maps.