Name
Abstract | ||
|---|---|---|
This paper explores the significance of phase information for speech emotion classification. The phase information is extracted from the discrete Fourier transform (DFT) spectrum. The phase of the pitch harmonic is used as a proposed feature for speech emotion classification. Pitch frequency varies with emotions, and due to this pitch harmonic also varies with different emotions. It is expected that the phase of the pitch harmonic contains emotion information. Significance of the harmonic phase is carried out by evaluating the mean and variance values for speech emotion classification. Support Vector Machine (SVM) classifier is used to evaluate the performance of the proposed feature. The performance is evaluated using EMODB database. The performance of the proposed feature is compared with the the linear prediction coefficients (LPC), mel frequency cepstral coefficients (MFCC) and Teager energy operator (TEO) based non-linear critical band TEO autocorrelation envelope (TEO-CB-Auto-Env) features. An average recognition rate of 73. 9% is achieved with the combination of the MFCC and proposed features. |
Year | Venue | Keywords |
|---|---|---|
2017 | National Conference on Communications NCC | Harmonic phase,speech emotion classification,pitch harmonic,K-S statistic |
Field | DocType | Citations |
Mel-frequency cepstrum,Pattern recognition,Computer science,Support vector machine,Emotion classification,Harmonic,Speech recognition,Linear prediction,Feature extraction,Artificial intelligence,Discrete Fourier transform,Autocorrelation | Conference | 0 |
PageRank | References | Authors |
0.34 | 15 | 2 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Deb, Suman | 1 | 24 | 4.59 |
| S. Dandapat | 2 | 261 | 28.51 |