Nonlinear Dynamic Feature Extraction Based on Phase Space Reconstruction for the Classification of Speech and Emotion

Abstract

Due to the shortcomings of linear feature parameters in speech signals, and the limitations of existing time- and frequency-domain attribute features in characterizing the integrity of the speech information, in this paper, we propose a nonlinear method for feature extraction based on the phase space reconstruction (PSR) theory. First, the speech signal was analyzed using a nonlinear dynamic model. Then, the model was used to reconstruct a one-dimensional time speech signal. Finally, nonlinear dynamic (NLD) features based on the reconstruction of the phase space were extracted as the new characteristic parameters. Then, the performance of NLD features was verified by comparing their recognition rates with those of other features (NLD features, prosodic features, and MFCC features). Finally, the Korean isolated words database, the Berlin emotional speech database, and the CASIA emotional speech database were chosen for validation. The effectiveness of the NLD features was tested using the Support Vector Machine classifier. The results show that NLD features not only have high recognition rate and excellent antinoise performance for speech recognition tasks but also can fully characterize the different emotions contained in speech signals.

Remote English Teaching Resource Sharing Based on Internet O2O Model

“…. , t n 􏼈 􏼉 represents the time series of the distribution of each data resource in a resource, F s represents the result of phase space reconstruction of each data resource structure, and ε represents the embedded dimension, the multidimensional state space of data resource structure is established based on the distribution point Q X of each data resource in high-dimensional space obtained in Section 3.4 and integrated into the theory of chaotic phase space reconstruction [35,36]:…”

Recurrence plot embeddings as short segment nonlinear features for multimodal speaker identification using air, bone and throat microphones

“…From the experiments so far (Tables 2, 3, 4,5,6,7,8,9,10,11), the evidence for RP embeddings being a standalone, nonlinear feature for speaker modeling is strong, and on par with the linear features (like MFCC). The improvement in performance in standalone RP based systems as the modes are combined, for unimodal, bimodal and trimodal systems is depicted in Fig.…”

“…These features however do not directly capture the non-linear dynamics of the speech signal, as they focus on extracting features related to the spectral envelop of the speech signal and inherently ignore the phase relationship between the frequency components 5 . Focusing on the piecewise linearity of the speech signal to extract features ignores crucial information on the non-linear dynamics, making the information being extracted incomplete 6 .…”