SigWavNet: Learning Multiresolution Signal Wavelet Network for Speech Emotion Recognition
Alaa Nfissi, Wassim Bouachir, Nizar Bouguila, Brian Mishara
TL;DR
SigWavNet introduces an end-to-end multiresolution framework for speech emotion recognition that learns representations directly from raw waveforms using a learnable fast discrete wavelet transform (FDWT) and a conjugate quadrature filter-based setup. The architecture combines learnable wavelet-like decomposition (Conv_h/Conv_g) with a learnable asymmetric hard-threshold (LAHT), followed by a 1D dilated CNN with spatial attention and a Bi-GRU with temporal attention, augmented by channel weighting and global average pooling. Empirical results on IEMOCAP and EMO-DB show state-of-the-art performance, with accuracy of 84.8% and 90.1% respectively, and ablation studies confirming the value of LAHT and full kernel learning. The work advances SER by offering a data-driven, noise-robust, multiresolution representation learned end-to-end, with code and model details available for reproducibility and further development.
Abstract
In the field of human-computer interaction and psychological assessment, speech emotion recognition (SER) plays an important role in deciphering emotional states from speech signals. Despite advancements, challenges persist due to system complexity, feature distinctiveness issues, and noise interference. This paper introduces a new end-to-end (E2E) deep learning multi-resolution framework for SER, addressing these limitations by extracting meaningful representations directly from raw waveform speech signals. By leveraging the properties of the fast discrete wavelet transform (FDWT), including the cascade algorithm, conjugate quadrature filter, and coefficient denoising, our approach introduces a learnable model for both wavelet bases and denoising through deep learning techniques. The framework incorporates an activation function for learnable asymmetric hard thresholding of wavelet coefficients. Our approach exploits the capabilities of wavelets for effective localization in both time and frequency domains. We then combine one-dimensional dilated convolutional neural networks (1D dilated CNN) with a spatial attention layer and bidirectional gated recurrent units (Bi-GRU) with a temporal attention layer to efficiently capture the nuanced spatial and temporal characteristics of emotional features. By handling variable-length speech without segmentation and eliminating the need for pre or post-processing, the proposed model outperformed state-of-the-art methods on IEMOCAP and EMO-DB datasets. The source code of this paper is shared on the Github repository: https://github.com/alaaNfissi/SigWavNet-Learning-Multiresolution-Signal-Wavelet-Network-for-Speech-Emotion-Recognition.