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

Analog "neuronal" networks in early vision.

C Koch, J Marroquin, A Yuille

    Proceedings of the National Academy of Sciences of the United States of America
    |June 1, 1986
    PubMed
    Summary

    This study presents a novel analog neural network approach for solving complex early vision problems, particularly those involving non-quadratic cost functions and discontinuities. This method enables efficient reconstruction of surfaces from sparse data, advancing artificial and biological vision systems.

    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

    Mapping the landscape of climate engineering.

    Philosophical transactions. Series A, Mathematical, physical, and engineering sciences·2014
    Same author

    An extremum principle for shape from contour.

    IEEE transactions on pattern analysis and machine intelligence·2011
    Same author

    Efficient multilevel brain tumor segmentation with integrated bayesian model classification.

    IEEE transactions on medical imaging·2008
    Same author

    Soil surface structure effect on isoproturon and diflufenican loss in runoff.

    Journal of environmental quality·2002
    Same author

    Brain areas specific for attentional load in a motion-tracking task.

    Journal of cognitive neuroscience·2002
    Same author

    Does a liquid supplement improve energy and protein consumption in nursing home residents?

    Geriatric nursing (New York, N.Y.)·2002

    Area of Science:

    • Computational Neuroscience
    • Computer Vision
    • Artificial Intelligence

    Background:

    • Early vision tasks often involve minimizing cost functions, with quadratic problems solvable by analog networks.
    • Non-quadratic cost functions, common with discontinuities, pose challenges for existing computational methods.
    • Nonlinear analog neural networks offer a potential solution for complex optimization problems.

    Purpose of the Study:

    • To generalize nonlinear analog neural networks for solving non-convex energy functionals in early vision.
    • To demonstrate an efficient computational strategy for early vision problems.
    • To reconstruct smooth surfaces from sparse data while preserving discontinuities.

    Main Methods:

    • Generalization of Hopfield and Tank's nonlinear analog neural networks.

    Related Experiment Videos

  • Implementation of a specific analog network for surface reconstruction.
  • Focus on minimizing non-quadratic cost functions in early vision.

    • Main Results:

    • Successfully adapted analog neural networks for non-convex optimization problems in early vision.
    • Developed an analog network capable of reconstructing surfaces from sparse data.
    • Preserved surface discontinuities during the reconstruction process.

    Conclusions:

    • Nonlinear analog neural networks provide an effective strategy for early vision problems with discontinuities.
    • The implemented analog network offers a novel computational approach for biological and artificial vision systems.
    • This research opens new avenues for real-time vision system development.