雑誌文献を検索します。書籍を検索する際には「書籍検索」を選択してください。

検索

書誌情報 詳細検索 by 医中誌

Japanese

Detecting Psychological Stress from Voice Signal Using the OpenEAR Software Takaya TAGUCHI 1,4 , Kiyotaka NEMOTO 1,2 , Hirokazu TACHIKAWA 1,2 , Toru NAGANO 3 , Ryuki TACHIBANA 3 , Masafumi NISHIMURA 5 , Tetsuaki ARAI 1,2 , Takashi ASADA 1,2 1Department of Psychiatry, Graduate School of Comprehensive Human Science, University of Tsukuba, Tsukuba, Japan 2Department of Psychiatry, Faculty of Medicine, University of Tsukuba 3IBM Research-Tokyo 4Yuhokai Kurita Hospital 5Graduate School of Informatics, Shizuoka University Keyword: Psychological stress , Voice analysis , OpenEAR , Audio features , MFCC , Mel-frequency cepstral coefficients pp.1027-1034
Published Date 2014/12/15
DOI https://doi.org/10.11477/mf.1405204067
  • Abstract
  • Look Inside
  • Reference

 In this study, we tried to detect psychological stress from voice samples using the OpenEAR software. One-hundred and six healthy subjects were randomly assigned to the control and the task group. First, State-Trait Anxiety Inventory (STAI), blood pressure, and pulse were measured and voices were also recorded for both groups. Then, the task group went through anagram task while the control group just read aloud the task phrases. Following the task, voices were recorded again. Voices as well as other parameters were compared between the two groups. We also investigated the correlation between audio features and difference value of STAI-S. As a result, we found that a part of Mel-frequency cepstral coefficients parameters showed significant change between control and task group, and pitch frequency as well as showed significant correlation with difference value of STAI-S. These results suggest that audio features extracted by OpenEAR might be indicators to detect psychological stress.


Copyright © 2014, Igaku-Shoin Ltd. All rights reserved.

基本情報

電子版ISSN 1882-126X 印刷版ISSN 0488-1281 医学書院

関連文献

もっと見る

文献を共有