Speech Coding and Compression Using Wavelets and Lateral Inhibitory Networks

Abstract

The purpose of this thesis is to introduce the concept of lateral inhibition as a generalized technique for compressing time/frequency representations of electromagnetic and acoustical signals, particularly speech. This requires at least a rudimentary treatment of the theory of frames-which generalizes most commonly known time/frequency distributions--the biology of hearing, and digital signal processing. As such, this material, along with the interrelationships of the disparate subjects, is presented in a tutorial style. This may leave the mathematician longing for more rigor, the neurophysiological psychologist longing for more substantive support of the hypotheses presented, and the engineer longing for a reprieve from the theoretical barrage. Despite the problems that arise when trying to appeal to too wide an audience, I hope this thesis is a cogent analysis of the compression of time/frequency distributions via lateral inhibitory networks. (js)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 1990
Accession Number
ADA230499

Entities

People

  • Richard Ricart

Organizations

  • Air Force Institute of Technology

Tags

DTIC Thesaurus Topics

  • Brain
  • Computer Programming
  • Crystal Structure
  • Data Compression
  • Digital Signal Processing
  • Ear
  • Electrical Engineering
  • Equations
  • Mathematical Filters
  • Recognition
  • Signal Processing
  • Software Development
  • Spectra
  • Speech Compression
  • Time Signals
  • Trees (Data Structures)
  • Wavelet Transforms

Readers

  • Radio communications and signal processing.
  • Speech Processing/Speech Recognition.
  • Theoretical Analysis.