A comparison of reconstructed phase spaces and cepstral coefficients for multi-band phoneme classification

Kevin M. Indrebo, Richard J. Povinelli, Michael T. Johnson

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

This paper examines the use of multi-band reconstructed phase spaces as models for phoneme classification. Sub-banding reconstructed phase spaces combines linear, frequency-based techniques with a nonlinear modeling approach to speech recognition. Experiments comparing the effects of filtering speech signals for both reconstructed phase space and traditional speech recognition approaches are presented. These experiments study the use of two non-overlapping subbands for isolated phoneme classification on the TIMIT corpus. It is shown that while classification accuracy using Mel frequency cepstral coefficients as features does not improve with sub-banding, the accuracy increases from 36.1% to 42.0% using sub-banded reconstructed phase spaces to model the phonemes.

Original languageEnglish
Title of host publication2004 7th International Conference on Signal Processing Proceedings, ICSP
Pages637-640
Number of pages4
StatePublished - 2004
Event2004 7th International Conference on Signal Processing Proceedings, ICSP - Beijing, China
Duration: Aug 31 2004Sep 4 2004

Publication series

Name2004 7th International Conference on Signal Processing Proceedings, ICSP

Conference

Conference2004 7th International Conference on Signal Processing Proceedings, ICSP
Country/TerritoryChina
CityBeijing
Period8/31/049/4/04

ASJC Scopus subject areas

  • Engineering (all)

Fingerprint

Dive into the research topics of 'A comparison of reconstructed phase spaces and cepstral coefficients for multi-band phoneme classification'. Together they form a unique fingerprint.

Cite this