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

Data driven image models through continuous joint alignment.

Erik G Learned-Miller1

  • 1Department of Computer Science, University of Massachusetts, Amherst, 140 Governor's Drive, Amherst, MA 01003, USA. elm@cs.umass.edu

IEEE Transactions on Pattern Analysis and Machine Intelligence
|February 14, 2006
PubMed
Summary
This summary is machine-generated.

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

Early diagnosis of acromegaly: computers vs clinicians.

Clinical endocrinology·2011
Same author

Many heads are better than one: jointly removing bias from multiple MRIs using nonparametric maximum likelihood.

Information processing in medical imaging : proceedings of the ... conference·2007
See all related articles

Congealing techniques model image classes by minimizing variations, enabling efficient learning for tasks like handwritten digit recognition and magnetic resonance image bias removal.

Area of Science:

  • Computer Vision
  • Machine Learning
  • Image Processing

Background:

  • Modeling image classes often requires handling variations like deformations or noise.
  • Existing methods may struggle with complex nuisance variables in image data.

Purpose of the Study:

  • To introduce congealing, a novel family of techniques for image class modeling.
  • To demonstrate the application of congealing for both generative modeling and nuisance variable removal.

Main Methods:

  • Congealing iteratively transforms images to minimize variations along specified axes.
  • The method models both latent images and nuisance variables, creating factorized generative models.
  • Nonparametric models are used for pixel intensity values, allowing cross-data learning.

Related Experiment Videos

Main Results:

  • Congealing successfully models image classes, demonstrated by building a handwritten digit classifier from minimal data.
  • The technique effectively removes bias fields from magnetic resonance images.
  • Experiments show robustness and consistency of the congealing method across various applications.

Conclusions:

  • Congealing offers a powerful and flexible approach to image class modeling and nuisance variable removal.
  • The technique facilitates efficient learning through knowledge sharing between tasks.
  • Congealing has broad applicability in computer vision and medical imaging.