Speech Emotion Recognition under White Noise

Abstract

Speaker's emotional states are recognized from speech signal with Additive white Gaussian noise (AWGN). The influence of white noise on a typical emotion recogniztion system is studied. The emotion classifier is implemented with Gaussian mixture model (GMM). A Chinese speech emotion database is used for training and testing, which includes nine emotion classes (e.g. happiness, sadness, anger, surprise, fear, anxiety, hesitation, confidence and neutral state). Two speech enhancement algorithms are introduced for improved emotion classification. In the experiments, the Gaussian mixture model is trained on the clean speech data, while tested under AWGN with various signal to noise ratios (SNRs). The emotion class model and the dimension space model are both adopted for the evaluation of the emotion recognition system. Regarding the emotion class model, the nine emotion classes are classified. Considering the dimension space model, the arousal dimension and the valence dimension are classified into positive regions or negative regions. The experimental results show that the speech enhancement algorithms constantly improve the performance of our emotion recognition system under various SNRs, and the positive emotions are more likely to be miss-classified as negative emotions under white noise environment.

An Investigation of a Feature-Level Fusion for Noisy Speech Emotion Recognition

“…The developed methods fall into two main categories. The first one includes speech enhancement or noise reduction techniques, either by means of speech sample reconstruction [7], noise compensation using histogram equalization [8], adaptive thresholding in the wavelet domain for noise cancellation [9], or spectral subtraction [10]. However, the downside of such methods is that they strongly rely on prior knowledge of noise, speech, or both, which limits their implementations.…”

A closed-form solution to the graph total variation problem for continuous emotion profiling in noisy environment

“…These noises severely degrade the performance of systems and consequently affect the user experience in real-life conditions Tawari and Trivedi (2010); Schuller et al (2006); Huang et al (2013); Schuller et al (2011). Therefore, a fundamental step in this work is to investigate environmental noise reduction via a novel closed-form solution to the graph signal theory-based method, to reduce noise at features level and improve the performance of emotion prediction.…”

“…In Zhang et al (2018) supervised single-channel technique is applied to speech dereverberation and denoising. In Tawari and Trivedi (2010), the authors utilized a speech enhancement technique based on the adaptive thresholding in the wavelet domain to address noises while in Huang et al (2013), the authors studied the influence of additive white Gaussian noise on speakers emotion states via a Gaussian mixture model. However, these methods are performed on discrete emotion conditions.…”

Noise and reverberation effects on depression detection from speech

“…Most studies use data collected in laboratory setups with little or no background noise or reverberation. The effect of noise on emotion prediction has been explored in [41,42,43], but little work has been done to study the effect of noise and/or reverberation on depression data.…”