Speech based detection of physiological stress using perceptually significant features

Yang Liu, Purdue University

Abstract

Verbal communication is a natural human communication pattern. The human vocal tract and articulators, which are biological organs with nonlinear characteristics, not only run under conscious control, but are also affected by gender and emotional state such as stress. Analysis of physiological stress based on features from speech parameters is preferable over other techniques because of its nonintrusive nature. In this thesis, heart rate (HR) at three different levels (low, medium and high heart rates) is used as a measure of stress and a common speech utterance in each case is employed to correlate stress with HR. Using a database consisting of utterances and their corresponding heart rates and other physiological measurements indicative of stress for the speakers, utterances of the word eye for a speaker at low, medium and high heart rates are isolated. From these short utterances, features based on auditorily significant areas are extracted for analysis. Spectral energy that is perceptually significant, i.e., above global masking threshold in each band, is used as a feature to classify speech under stress as indicated by measured heart rate. For comparison, the results of employing mel-frequency cepstral coefficients for classifying stress are also shown.

Degree

M.S.E.

Advisors

Gopalan, Purdue University.

Subject Area

Electrical engineering|Computer science|Physiological psychology

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS