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

Sequential and parallel image restoration: neural network implementations.

M T Figueiredo1, J N Leitao

  • 1Dept. de Engenharia Electrotecnica e de Comput., Inst. Superior Tecnico, Lisbon.

IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society
|January 1, 1994
PubMed
Summary

This study introduces novel neural network algorithms for image restoration, improving upon existing methods for blurred and noisy images. The new approach ensures algorithm convergence for enhanced image quality.

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

Absolute phase image reconstruction: a stochastic nonlinear filtering approach.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2008
Same author

Unsupervised contour representation and estimation using B-splines and a minimum description length criterion.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2008
Same author

Wavelet-based image estimation: an empirical Bayes approach using Jeffrey's noninformative prior.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2008
Same author

Hormonal receptors in mammary carcinoma: comparison between quantitative and qualitative methods.

Sao Paulo medical journal = Revista paulista de medicina·1997
Same author

Unsupervised image restoration and edge location using compound Gauss-Markov random fields and the MDL principle.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·1997
Same author

Wall position and thickness estimation from sequences of echocardiographic images.

IEEE transactions on medical imaging·1996

Area of Science:

  • Artificial Intelligence
  • Computer Vision
  • Computational Neuroscience

Background:

  • Image restoration addresses degradation from blur and noise.
  • Neural networks commonly use energy minimization for optimization.
  • Existing methods map optimization to network energy, limiting performance.

Purpose of the Study:

  • Propose novel neural network implementations for image restoration.
  • Develop sequential and parallel algorithms for improved performance.
  • Address limitations of current energy-minimization-based neural network strategies.

Main Methods:

  • Developed neural implementations of iterative minimization algorithms.
  • Utilized modified Hopfield networks with graded elements.

Related Experiment Videos

  • Investigated both sequential and parallel updating schedules.
  • Considered standard Hopfield networks for comparison.

    • Main Results:

    • Algorithms were proven to converge for image restoration tasks.
    • Demonstrated robustness with respect to finite numerical precision.
    • Presented successful examples using real-world degraded images.

    Conclusions:

    • Novel neural network algorithms offer effective image restoration.
    • Iterative minimization approaches provide a viable alternative to energy mapping.
    • The proposed methods enhance image quality for degraded inputs.