Seeing Through the Conversation: Audio-Visual Speech Separation based on Diffusion Model

TL;DR

This paper introduces AVDiffuSS, a diffusion-based audio-visual speech separation model that effectively fuses modalities using cross-attention, achieving state-of-the-art naturalness and separation quality on benchmark datasets.

Contribution

The paper presents a novel diffusion-based framework with a cross-attention fusion mechanism for improved audio-visual speech separation.

Findings

Achieves state-of-the-art results on VoxCeleb2 and LRS3 datasets.

Produces speech with significantly improved naturalness.

Maintains high temporal resolution with efficient computation.

Abstract

The objective of this work is to extract target speaker's voice from a mixture of voices using visual cues. Existing works on audio-visual speech separation have demonstrated their performance with promising intelligibility, but maintaining naturalness remains a challenge. To address this issue, we propose AVDiffuSS, an audio-visual speech separation model based on a diffusion mechanism known for its capability in generating natural samples. For an effective fusion of the two modalities for diffusion, we also propose a cross-attention-based feature fusion mechanism. This mechanism is specifically tailored for the speech domain to integrate the phonetic information from audio-visual correspondence in speech generation. In this way, the fusion process maintains the high temporal resolution of the features, without excessive computational requirements. We demonstrate that the proposed framework achieves state-of-the-art results on two benchmarks, including VoxCeleb2 and LRS3, producing speech with notably better naturalness.