Evaluation the Performance of Different Parameters To RecognizeEmotions from Speech
Horkous Houari, Guerti Mhania
Abstract
Recognition emotion in speech signal has attracted much attention and plays an important role in affect computing, artificial ch The aim of speech emotion recognition system is to extract the information from the spee intelligence and signal processing areas. emotional speech corpus in Algerian Dialect is used for In this work signal and identify the emotional state of a human being. (pitch, c prosodi in our study are the parameters The selected to analyze the emotions of fear, anger and neutral. parameters extraction . Frequency Cepstral Coefficients Mel parameters MFCCs ( cepstral and intensity and duration), the unvoiced frames, jitter, shimmer us to lead obtained results The Nearest Neighbor). Neighbor).--(K The system of recognition is based on the method of classification KNN observe that combined prosodies, jitter, shimmer, unvoiced frames and MFCCs parameters give recognition rate important (84.02
Details
| Original language | English |
|---|---|
| Event | International PluridisciplinaryPhD Meeting (IPPM’20) |
| Publication status | Published — February 2020 |