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

CNN: a speaker recognition system using a cascaded neural network

M Zaki, A Ghalwash, A A Elkouny

    International Journal of Neural Systems
    |May 1, 1996
    PubMed
    Summary
    This summary is machine-generated.

    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

    Mass online training of health care workers during COVID-19: approach, impact, and outcomes for over 10,000 health care providers.

    Public health·2024
    Same author

    Human Pose Estimation for Clinical Analysis of Gait Pathologies.

    Bioinformatics and biology insights·2024
    Same author

    New Au/chitosan nanocomposite modified carbon paste sensor for voltammetric detection of nicotine.

    Scientific reports·2023
    Same author

    Late Dacron Patch Inflammatory Reaction after Carotid Endarterectomy.

    European journal of vascular and endovascular surgery : the official journal of the European Society for Vascular Surgery·2017
    Same author

    A randomized cross-over study comparing the performance of HD integra™ central concentrate system versus pre-produced concentrate in hemodialysis.

    BMC nephrology·2017
    Same author

    The Role of Sequential Pneumatic Compression in Limb Salvage in Non-reconstructable Critical Limb Ischemia.

    European journal of vascular and endovascular surgery : the official journal of the European Society for Vascular Surgery·2016

    This study combines supervised and unsupervised neural networks for speaker recognition, creating a cascaded model. The neural network approach significantly outperforms conventional methods, especially with noisy speech patterns.

    Area of Science:

    • Speech processing
    • Artificial intelligence
    • Machine learning

    Background:

    • Conventional speaker recognition relies on distance metrics sensitive to intraspeaker variance.
    • Neural networks offer an alternative approach to speaker recognition.
    • Combining supervised and unsupervised learning can enhance recognition performance.

    Discussion:

    • A novel cascaded model integrates supervised (pattern association) and unsupervised (winner-take-all) neural network models.
    • The supervised model acts as a pre-filtration stage to handle noisy patterns.
    • Conventional methods using distance metrics are presented for performance comparison.

    Key Insights:

    • The cascaded neural network model demonstrates superior performance compared to conventional methods.

    Related Experiment Videos

  • The system exhibits smoother performance degradation with noisy speech patterns.
  • Noise-free patterns achieve higher recognition accuracy with the neural network approach.

    • Outlook:

    • Further research can explore advanced feature extraction techniques for speaker recognition.
    • Investigating different neural network architectures may yield further performance improvements.
    • Real-world deployment of cascaded models in secure communication systems is a potential future direction.