Lite Audio-Visual Speech Enhancement

TL;DR

This paper introduces a lightweight audio-visual speech enhancement system that improves denoising performance while reducing computational costs and privacy concerns through visual data compression and model simplification.

Contribution

The paper proposes a novel Lite AVSE system that incorporates visual data compression and removes visual feature extraction to enhance efficiency and privacy.

Findings

LAVSE outperforms audio-only systems with similar parameters.

Two visual data compression techniques are effective.

The system achieves better performance with reduced computational costs.

Abstract

Previous studies have confirmed the effectiveness of incorporating visual information into speech enhancement (SE) systems. Despite improved denoising performance, two problems may be encountered when implementing an audio-visual SE (AVSE) system: (1) additional processing costs are incurred to incorporate visual input and (2) the use of face or lip images may cause privacy problems. In this study, we propose a Lite AVSE (LAVSE) system to address these problems. The system includes two visual data compression techniques and removes the visual feature extraction network from the training model, yielding better online computation efficiency. Our experimental results indicate that the proposed LAVSE system can provide notably better performance than an audio-only SE system with a similar number of model parameters. In addition, the experimental results confirm the effectiveness of the two techniques for visual data compression.