Abstract
|
|
---|---|
New tools based on speech analysis can improve and accelerate diagnosis of Parkinson's disease. In this work, the use of some specific segments of speech, around the so called Acoustic Landamarks, are used with different families of features such as acoustic cues or Rasta-PLP and GMM-UBM-Blend classification methods to detect Parkinson's disease. Results of 87% are obtained. Burst segmentes provide the most relevant information when detecting Parkinson's disease whjile GMM-UMB-Blend is revealed as a promising techinique when using small databases and segmented speech. | |
International
|
Si |
Congress
|
10th International Workshop Models and Analysis of Vocal Emissions for Biomedical Applications |
|
960 |
Place
|
|
Reviewers
|
Si |
ISBN/ISSN
|
978-88-6453-607-1 |
|
|
Start Date
|
13/12/2017 |
End Date
|
15/12/2017 |
From page
|
73 |
To page
|
76 |
|
Models and Analysis of Vocal Emissions for Biomedical Applications: 10th Internation Workshop |