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

High-order contrasts for independent component analysis

J F Cardoso1

  • 1Ecole Nationale Supérieure des Télecommunications, ENST / SIG, 46 rue Barrault, 75634 Paris Cedex 13, France. cardoso@sig.enst.fr

Neural Computation
|February 9, 1999
PubMed
Summary

This study explores advanced methods for measuring independence in independent component analysis (ICA). Jacobi algorithms offer an efficient approach for optimizing these measures, outperforming gradient methods on biomedical data.

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

Spiramyin-loaded PLGA implants for the treatment of ocular toxoplasmosis: development, characterization, biocompatibility, and anti-toxoplasma activity.

Die Pharmazie·2021
Same author

Full-length genomic and molecular characterization of Canine parvovirus in dogs from North of Brazil.

Genetics and molecular research : GMR·2017
Same author

Optimization of dengue virus genome assembling using GSFLX 454 pyrosequencing data: evaluation of assembling strategies.

Genetics and molecular research : GMR·2012
Same author

Sedation of children undergoing MRI--a risky business!

South African medical journal = Suid-Afrikaanse tydskrif vir geneeskunde·2004
Same author

Ultrasonic signal processing for in vivo attenuation measurement: short time Fourier analysis.

Ultrasonic imaging·1983
Same author

[Evaluation of the needs of the official services of public health in São Paulo municipality. At present and in the future].

Revista de saude publica·1969

Area of Science:

  • Computational statistics
  • Signal processing
  • Machine learning

Background:

  • Independent Component Analysis (ICA) is a blind source separation technique.
  • Measuring independence is crucial for ICA performance.
  • Optimization algorithms are needed for efficient ICA.

Purpose of the Study:

  • To investigate high-order measures of independence for ICA.
  • To evaluate Jacobi algorithms for optimizing these measures.
  • To compare Jacobi algorithms with gradient-based methods.

Main Methods:

  • Exploration of high-order independence measures.
  • Application and discussion of Jacobi algorithms for optimization.
  • Algorithmic comparison with gradient-based techniques.

Related Experiment Videos

  • Validation on biomedical datasets.

    • Main Results:

    • Jacobi algorithms provide an effective optimization strategy for high-order independence measures.
    • Proposed Jacobi-based approaches demonstrate competitive or superior performance compared to gradient methods.
    • Successful application and validation on real-world biomedical data.

    Conclusions:

    • High-order independence measures enhance ICA.
    • Jacobi algorithms are efficient and effective for ICA optimization.
    • The proposed methods show promise for biomedical signal processing applications.