Features extraction for speech emotion

Abstract

In this paper the speech emotion verification using two most popular methods in speech processing and analysis based on the Mel-Frequency Cepstral Coefficient (MFCC) and the Gaussian Mixture Model (GMM) were proposed and analyzed. In both cases, features for the speech emotion were extracted using the Short Time Fourier Transform (STFT) and Short Time Histogram (STH) for MFCC and GMM respectively. The performance of the speech emotion verification is measured based on three neural network (NN) and fuzzy neural network (FNN) architectures; namely: Multi Layer Perceptron (MLP), Adaptive Neuro Fuzzy Inference System (ANFIS) and Generic Self-organizing Fuzzy Neural Network (GenSoFNN). Results obtained from the experiments using real audio clips from movies and television sitcoms show the potential of using the proposed features extraction methods for real time application due to its reasonable accuracy and fast training time. This may lead us to the practical usage if the emotion verifier can be embedded in real time applications especially for personal digital assistance (PDA) or smart-phones.

Emotion recognition from speech: a review

“…A four stage ergodic hidden Markov model (HMM) is used as a classifier to accomplish this task. Performance of LFPC parameters is compared with conventional LPCC and MFCC features, and observed that LFPCs perform slightly better (Williams and Stevens 1981;Kamaruddin and Wahab 2009). The MFCC features extracted from lower frequency components (20 Hz to 300 Hz) of speech signal are proposed to model pitch variation.…”

“…Prosodic Yildirim et al (2004), Ververidis et al (2004), and McGilloway et al (2000) Spectral Lee et al (2001), Yildirim et al (2004), Ververidis et al (2004), and Lee and Narayanan (2005) 07 Hidden Markov models (HMM) Prosodic Fernandez and Picard (2003), Zhou et al (2001), Nwe et al (2003), and Bitouk et al (2010)SpectralWilliams and Stevens (1981),Zhou et al (2001),Nwe et al (2003), andKamaruddin and Wahab (2009) …”

Measuring Customer Perceptions Index Using CMAC Speech Emotion Mapping

“…Further explanations of the datasets are reported in [9,10]. In this paper, only two emotion of happiness and sadness will be used to discriminate respondent excitation level.…”

“…The NTU-American [9,10] and The Berlin Emo-db [11] datasets representing American and European cultural influence respectively. Further explanations of the datasets are reported in [9,10].…”

“…In addition, neutral is also taken into consideration that serves as the emotionless state. The NTU_American [9,10] and Berlin Emo-db [11] data corpus will be used to test the feasibility of our proposed approach.…”