Robust Unsupervised Audio-visual Speech Enhancement Using a Mixture of Variational Autoencoders

TL;DR

This paper introduces a robust unsupervised audio-visual speech enhancement technique that dynamically switches between audio-only and audio-visual VAEs to handle noisy visual data, improving speech quality in challenging conditions.

Contribution

It proposes a novel mixture model combining trained audio-only and audio-visual VAEs with a variational EM algorithm for robust speech enhancement.

Findings

The method effectively skips noisy visual frames by switching models.

Experimental results demonstrate improved speech enhancement performance.

The approach outperforms existing models in noisy visual scenarios.

Abstract

Recently, an audio-visual speech generative model based on variational autoencoder (VAE) has been proposed, which is combined with a nonnegative matrix factorization (NMF) model for noise variance to perform unsupervised speech enhancement. When visual data is clean, speech enhancement with audio-visual VAE shows a better performance than with audio-only VAE, which is trained on audio-only data. However, audio-visual VAE is not robust against noisy visual data, e.g., when for some video frames, speaker face is not frontal or lips region is occluded. In this paper, we propose a robust unsupervised audio-visual speech enhancement method based on a per-frame VAE mixture model. This mixture model consists of a trained audio-only VAE and a trained audio-visual VAE. The motivation is to skip noisy visual frames by switching to the audio-only VAE model. We present a variational expectation-maximization method to estimate the parameters of the model. Experiments show the promising performance of the proposed method.