Supervised and Unsupervised Feature Extraction from a Cochlear Model for Speech Recognition

Abstract

We explore the application of a novel classification method that combines supervised and unsupervised training, and compare its performance to various more classical methods. We first construct a detailed high dimensional representation of the speech signal using Lyon's cochlear model and then optimally reproduce its dimensionality. The resulting low dimensional projection retains the information needed for robust speech recognition.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 23, 1992
Accession Number
ADA261523

Entities

People

  • G. Tajchman
  • N. Intrator

Organizations

  • Brown University

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Auditory Nerve
  • Automated Speech Recognition
  • Dimensionality Reduction
  • Feature Extraction
  • Frequency
  • Information Science
  • Machine Learning
  • Membranes
  • Military Research
  • Neural Networks
  • Preprocessing
  • Recognition
  • Signal Processing
  • Supervised Machine Learning
  • Test Sets
  • United States
  • Unsupervised Machine Learning

Fields of Study

  • Computer science

Readers

  • Auditory Neuroscience/Auditory Physiology.
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML
  • AI & ML - Machine Learning Algorithms