2016 Fiscal Year Final Research Report
A detection method using phonetic segments for the influence of fatigue on speech
| Project/Area Number |
25330172
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Multi-year Fund |
|---|
| Section | 一般 |
|---|
| Research Field |
Cognitive science
|
|---|
| Research Institution | University of Shizuoka |
|---|
Principal Investigator |
|
|---|
| Research Collaborator |
IMOTO Tomoaki 静岡県立大学, 経営情報学部, 助教 (20749296)
|
|---|
| Project Period (FY) |
2013-04-01 – 2017-03-31
|
| Keywords | 音声セグメント / 疲労 / 心拍変動解析 / フリッカー値 / 唾液アミラーゼ / 自覚症しらべ / 基本周波数 / ストレス |
|---|
| Outline of Final Research Achievements |
To estimate degree of fatigue using speech, changes on speech caused by stressor were investigated. A correlation analysis and a multiple regression analysis were conducted between each fatigue degree index and speech parameter. Variable selection was performed in multiple regression analysis to predict the LF/HF. As a result, four phonetic segment labels in the utterance "ah" and in the utterance "oh", and F0 in the utterance "start" were selected as speech parameters. The correlation coefficient between the estimated and actual values became 0.98 in the arithmetic operation stressor and 0.90 in the anaerobic stressor. And it was shown that the speech parameters may substitute the LF/HF.
|
| Free Research Field |
音声情報処理
|
