Descripción
|
|
---|---|
Disordered voices are frequently assessed by speech pathologists using acoustic perceptual evaluations. This might lead to problems due to the subjective nature of the process and due to the in uence of external factors which compromise the quality of the assessment. In order to increase the reliability of the evaluations the design of new indicator parameters obtained from voice signal processing is desirable. With that in mind, this paper presents an automatic evaluation system which emulates perceptual assessments of the roughness level in human voice. Two parameterization methods are used: complexity, which has already been used successfully in previous works, and modulation spectra. For the latter, a new group of parameters has been proposed as Low Modulation Ratio (LMR), Contrast (MSW) and Homogeneity (MSH). The tested methodology also employs PCA and LDA to reduce the dimensionality of the feature space, and GMM classi?ers for evaluating the ability of the proposed features on distinguishing the dierent roughness levels. An e?ciency of 82% and a Cohen's Kappa Index of 0:73 is obtained using the modulation spectra parameters, while the complexity parameters performed 73% and 0:58 respectively. The obtained results indicate the usefulness of the proposed modulation spectra features for the automatic evaluation of voice roughness which can derive in new parameters to be useful for clinicians. | |
Internacional
|
No |
Nombre congreso
|
VIII Jornadas en Tecnologías del Habla and IV Iberian SLTech Workshop - IBERSPEECH 2014 |
Tipo de participación
|
970 |
Lugar del congreso
|
|
Revisores
|
Si |
ISBN o ISSN
|
978-84-617-2862-6 |
DOI
|
|
Fecha inicio congreso
|
19/11/2014 |
Fecha fin congreso
|
21/11/2104 |
Desde la página
|
71 |
Hasta la página
|
80 |
Título de las actas
|
Conference Prodeedings IberSpeech 2014 |