Who said that?: Audio-visual speaker diarisation of real-world meetings

TL;DR

This paper presents a novel audio-visual diarisation method for real-world meetings that combines video and multi-channel audio to accurately determine who spoke when, outperforming existing approaches.

Contribution

It introduces an iterative approach that leverages audio-visual correspondence for speaker enrollment and active speaker detection, with enhanced performance using beamforming.

Findings

Strong results on real-world meeting data

Outperforms all comparable methods on the AMI corpus

Beamforming improves multi-channel audio diarisation

Abstract

The goal of this work is to determine 'who spoke when' in real-world meetings. The method takes surround-view video and single or multi-channel audio as inputs, and generates robust diarisation outputs. To achieve this, we propose a novel iterative approach that first enrolls speaker models using audio-visual correspondence, then uses the enrolled models together with the visual information to determine the active speaker. We show strong quantitative and qualitative performance on a dataset of real-world meetings. The method is also evaluated on the public AMI meeting corpus, on which we demonstrate results that exceed all comparable methods. We also show that beamforming can be used together with the video to further improve the performance when multi-channel audio is available.