2011 Fiscal Year Final Research Report
Forensic speaker recognition with global classification of speech
| Project/Area Number |
21710174
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
Social systems engineering/Safety system
|
|---|
| Research Institution | National Research Institute of Police Science |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2009 – 2011
|
| Keywords | 話者認識 / 法科学 |
|---|
| Research Abstract |
This study aimed to develop a method for forensic speaker recognition, using global classification of speech. To confirm whether speakers were classified to a few classes, we conducted cluster analysis in which speakers were classified according to their formant frequencies. Formant frequencies were extracted from vowels of 319 female and 313 male speakers by ARX-based speech analysis method. In addition, a novel model for predicting formant frequency from speech signals was developed, which can efficiently represent speech information.
|
