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

A network for recursive extraction of canonical coordinates.

Ali Pezeshki1, Mahmood R Azimi-Sadjadi, Louis L Scharf

  • 1Department of Electrical and Computer Engineering, Colorado State University, Fort Collins, CO 80523-1373, USA. ali@engr.colostate.edu

Neural Networks : the Official Journal of the International Neural Network Society
|July 10, 2003
PubMed
Summary

A novel network structure efficiently extracts canonical coordinates from data channels using linear subnetworks and lateral connections. This design enables scalable coordinate extraction without retraining previous network components.

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

Misspecified Cramér-Rao lower bound with Poisson statistics and its application in localization microscopy.

Optics letters·2025
Same author

Statistical estimation theory detection limits for label-free imaging.

Journal of biomedical optics·2024
Same author

Phylogenetic analysis of Acanthamoeba isolated from soil samples and nasal cavities of patients with malignancy: a public health concern in the northwest of Iran.

Transactions of the Royal Society of Tropical Medicine and Hygiene·2024
Same author

Super-resolution computational saturated absorption microscopy.

Journal of the Optical Society of America. A, Optics, image science, and vision·2023
Same author

Fisher information and the Cramér-Rao lower bound in single-pixel localization microscopy with spatiotemporally modulated illumination.

Journal of the Optical Society of America. A, Optics, image science, and vision·2023
Same author

Detection of Toxocara species larvae in four Iranian free-range broiler farms.

BMC veterinary research·2022

Area of Science:

  • Machine Learning
  • Data Analysis
  • Network Architecture

Background:

  • Canonical coordinate decomposition is crucial for dimensionality reduction and feature extraction.
  • Existing methods may require retraining when adding new components, limiting scalability.
  • Efficient and adaptable methods for coordinate extraction are needed.

Purpose of the Study:

  • To present a new network structure for canonical coordinate decomposition.
  • To enable scalable extraction of canonical coordinates from multiple data channels.
  • To develop a method that avoids retraining when adding new extraction nodes.

Main Methods:

  • A network composed of two single-layer linear subnetworks is proposed.
  • Canonical coordinates are extracted using a hierarchical set of lateral connections.

Related Experiment Videos

  • A stochastic gradient descent learning algorithm trains the network's connection weights.
  • A deflation process is implemented within subnetworks to refine coordinate extraction.

    • Main Results:

    • The network successfully extracts canonical coordinates from two data channels.
    • The hierarchical lateral connections facilitate a deflation process.
    • The proposed structure allows for the addition of new nodes without retraining.
    • Performance was validated using a synthesized data set.

    Conclusions:

    • The presented network structure offers an efficient and scalable approach to canonical coordinate decomposition.
    • The architecture's ability to add nodes without retraining enhances its practical applicability.
    • This method provides a foundation for more complex data analysis tasks requiring sequential feature extraction.