International Journal of Computer Theory and Engineering

Abstract—Evaluation of speakers who are high-risk suicidal compared to those with less clinical depression are critical when the syndrome underlying a patient’s abnormal behaviour is diagnosed without expertise. This study describes a way to classify the speech samples collected from groups of depressive and suicidal speakers by employing the speech processing technique in data analysis. First, the Glottal Spectral Slope (GSS) and Mel-Frequency Cepstral Coefficients (MFCC) were computationally estimated from the voiced segments detected from the categorized speech sample database. Second, the pairwise classification was then made on the combination of those extracted vocal features respectively corresponding to the frequency response of the source and the filter in speech production system model.
The procedure of this research was carried out in order to investigate the discriminative property of the focused vocal parameters mainly between depressed speakers and high-risk suicidal speaker groups. The result revealed that MFCC and GSS parameters are slightly high effective in term of vocal indicator corresponding to severe depression with fairly high performance in between-group separation.

Index Terms—Depression, glottal spectral slope, MFCC, speech.

Thaweewong Akkaralaertsest is with Division of Electronics and Telecommunication Engineering, Faculty of Engineering, Rajamangala University of Technology KrungThep, Bangkok, Thailand (e-mail: thaweewong.a@rmutk.ac.th).
Thaweesak Yingthawornsuk is with Media Technology, King Mongkut's University of Technology Thonburi – Bang Khuntien Campus, Bang Khuntien, Bangkok, 10150 Thailand (e-mail: thaweesak.yin@kmutt.ac.th).

[PDF]

Cite:Thaweewong Akkaralaertsest and Thaweesak Yingthawornsuk, "Comparative Analysis of Vocal Characteristics in Speakers with Depression and High-Risk Suicide," International Journal of Computer Theory and Engineering vol. 7, no. 6, pp. 448-452, 2015.