Audio-Visual Speech Enhancement with Score-Based Generative Models
This work addresses speech enhancement for noisy audio scenarios by leveraging visual cues, offering an incremental improvement over audio-only methods.
The paper tackled speech enhancement by integrating visual lipreading features into a diffusion model, resulting in improved speech quality and reduced phonetic confusions, with a noticeable decrease in word error rate for automatic speech recognition, especially at low signal-to-noise ratios.
This paper introduces an audio-visual speech enhancement system that leverages score-based generative models, also known as diffusion models, conditioned on visual information. In particular, we exploit audio-visual embeddings obtained from a self-super\-vised learning model that has been fine-tuned on lipreading. The layer-wise features of its transformer-based encoder are aggregated, time-aligned, and incorporated into the noise conditional score network. Experimental evaluations show that the proposed audio-visual speech enhancement system yields improved speech quality and reduces generative artifacts such as phonetic confusions with respect to the audio-only equivalent. The latter is supported by the word error rate of a downstream automatic speech recognition model, which decreases noticeably, especially at low input signal-to-noise ratios.