Teager Energy-Autocorrelation Envelope for Stressed Speech Emotion Recognition with Spectral Features: A Multi-database Analysis

Abstract

A new feature extraction technique using Teager Energy Operator is proposed for the detection of stressed sentiments as Teager Energy-Autocorrelation Envelope. TEO is basically designed for increasing the energies of the stressed speech signal whose energies are reduced during the speeches production process and hence, used in these analysis. A stressed speech emotion recognition system is developed employing TEO-Auto-Env and Spectral feature combination for detecting the emotions. Mel frequency cepstral coefficients, linear prediction cepstral coefficients, and relative spectra—perceptual linear prediction are the spectral properties studied. EMO-DB (German), EMOVO (Italian), IITKGP (Telugu) and EMA (English) databases are used in this analysis. The classification of the emotions is carried out using the k-Nearest Neighborhood classifiers for gender-dependent and speaker-independent cases. The proposed SSER system provided improved precision comparison to the previous ones. The greatest classification precision is obtained using the characteristic combination of TEO-Auto-Env, MFCC and LPCC features with 91.4% (SI), 91.4% (GD-Male) and 93.1%(GD-female) for EMO-DB, 68.5% (SI), 68.5% (GD-Male) and 74.6% (GD-female) for EMOVO, 90.6%(SI), 91% (GD-Male) and 92.3% (GD-female) for EMA, and 95.1% (GD-female) for IITKGP female database.