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

Blur identification by multilayer neural network based on multivalued neurons.

Igor Aizenberg1, Dmitriy V Paliy, Jacek M Zurada

  • 1Texas A&M University-Texarkana, Texarkana, TX 75505 USA. igor.aizenberg@tamut.edu

IEEE Transactions on Neural Networks
|May 10, 2008
PubMed
Summary

Multilayer multivalued neuron networks (MLMVN) offer superior, flexible, and faster image deblurring by accurately identifying point spread function types and parameters. This derivative-free approach excels in complex classification tasks.

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

Deep Learning for Image Watermarking: A Comprehensive Review and Analysis of Techniques, Challenges, and Applications.

Sensors (Basel, Switzerland)·2026
Same author

Bidirectional Multiscale Efficient Dilated Convolutional Recurrent Neural Network Improved by Swarm Intelligence Optimization.

IEEE transactions on neural networks and learning systems·2025
Same author

Patient-Specific Modeling and Model Predictive Control Approach to Personalized Optimal Anemia Management.

Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference·2023
Same author

Convergence Analysis of Online Gradient Method for High-Order Neural Networks and Their Sparse Optimization.

IEEE transactions on neural networks and learning systems·2023
Same author

Control-Relevant Adaptive Personalized Modeling From Limited Clinical Data for Precise Warfarin Management.

IEEE open journal of engineering in medicine and biology·2023
Same author

Adaptive Individualized Drug-Dose Response Modeling from a Limited Clinical Data: Case of Warfarin Management.

Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference·2021

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Computer Vision

Background:

  • Traditional feedforward neural networks and kernel-based networks have limitations in modeling complex mappings.
  • Accurate identification of the point spread function (PSF) is critical for effective image deblurring.

Purpose of the Study:

  • To introduce and evaluate a multilayer neural network based on multivalued neurons (MLMVN) for image deblurring applications.
  • To demonstrate the MLMVN's capability in identifying both the type and parameters of the point spread function.

Main Methods:

  • Implementation of a multilayer neural network with multivalued neurons (MLMVN).
  • Utilizing a derivative-free backpropagation learning algorithm for network training.
  • Applying the MLMVN to classify and determine parameters of the point spread function in simulated image data.

Related Experiment Videos

Main Results:

  • The MLMVN demonstrated superior functionality compared to traditional feedforward and kernel-based networks.
  • The network exhibited higher flexibility and faster adaptation for target mapping.
  • Simulation results confirmed the high efficiency of the MLMVN in identifying PSF type and parameters, crucial for image deblurring.
  • The MLMVN proved effective for classification problems, particularly multiclass scenarios.

Conclusions:

  • The MLMVN is a powerful and efficient tool for image deblurring through precise point spread function identification.
  • The network's advanced features enable modeling complex problems with simpler architectures.
  • MLMVN shows significant potential for various classification tasks, especially in complex, multiclass environments.