Feature Extraction and Decision Processes in the Classification of Amplitude Modulated Noise Patterns.

Abstract

The relation between the perceptual features identified in a multidimensional scaling (MDS) analysis and the decision stage of the auditory classification process was investigated in four experiments based upon a set of sixteen complex acoustic patterns. The sounds consisted of broad-band white noise, amplitude modulated by sawtooth waves of varying frequency and attack. Overall, results of the four experiments indicated that listeners employed an optimum-processor strategy to determine the relative importance of each feature in the decision process. The findings indicate that any theoretical treatment of auditory pattern recognition must address the interaction of the feature extraction and decision processes.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 1978
Accession Number
ADA058424

Entities

People

  • Donald C. Burgy
  • James A. Ballas
  • James H. Howard Jr.

Organizations

  • The Catholic University of America

Tags

Communities of Interest

  • Biomedical
  • Ground and Sea Platforms
  • Human Systems

DTIC Thesaurus Topics

  • Behavioral Sciences
  • Biological Sciences
  • Biomedical Research
  • Biophysics
  • Data Science
  • Engineering
  • Feature Extraction
  • Frequency
  • Human Factors Engineering
  • Industrial Engineering
  • Military Research
  • New York
  • Operations Research
  • Pattern Recognition
  • Probability
  • Psychology
  • Systems Engineering

Readers

  • Radar Systems Engineering.
  • Speech Processing/Speech Recognition.
  • Systems Analysis and Design

Technology Areas

  • AI & ML