Integrating Statistical Uncertainty Into Neural Network-based Speech Enhancement
2022 · Huajian Fang, Tal Peer, Stefan Wermter, et al.
Abstract
Speech enhancement in the time-frequency domain is often performed by estimating a multiplicative mask to extract clean speech. However, most neural network-based methods perform point estimation, i.e., their output consists of a single mask. In this paper, we study the benefits of modeling uncertainty in neural network-based speech enhancement. For this, our neural network is trained to map a noisy spectrogram to the Wiener filter and its associated variance, which quantifies uncertainty, based on the maximum a posteriori (MAP) inference of spectral coefficients. By estimating the distribution instead of the point estimate, one can model the uncertainty associated with each estimate. We further propose to use the estimated Wiener filter and its uncertainty to build an approximate MAP (A-MAP) estimator of spectral magnitudes, which in turn is combined with the MAP inference of spectral coefficients to form a hybrid loss function to jointly reinforce the estimation. Experimental results
Authors
(none)
Tags
Stats
Related papers
- Leveraging Heteroscedastic Uncertainty In Learning Complex Spectral Mapping For Single-channel Speech Enhancement (2022)3.58
- Statistical Speech Enhancement Based On Probabilistic Integration Of Variational Autoencoder And Non-negative Matrix Factorization (2017)15.00
- Neural Kalman Filtering For Speech Enhancement (2020)2.26
- Consistency-aware Multi-channel Speech Enhancement Using Deep Neural Networks (2020)0.00
- Spectral Masking With Explicit Time-context Windowing For Neural Network-based Monaural Speech Enhancement (2024)3.58
- An Improved Uncertainty Propagation Method For Robust I-vector Based Speaker Recognition (2019)5.84
- Single-channel Speech Enhancement With Deep Complex U-networks And Probabilistic Latent Space Models (2023)5.24
- Towards Speech Enhancement Using A Variational U-net Architecture (2020)7.81