Best of Both Worlds: Multi-task Audio-Visual Automatic Speech Recognition and Active Speaker Detection

TL;DR

This paper introduces a multi-task model that jointly performs audio-visual speech recognition and active speaker detection, significantly improving ASD accuracy while enhancing ASR performance in multi-person scenarios.

Contribution

It presents a novel joint training approach that reduces active speaker detection errors by 25% and boosts speech recognition accuracy compared to single-task models.

Findings

Active speaker detection accuracy improved by ~25%.

Speech recognition performance increased in multi-person settings.

Joint training benefits both tasks simultaneously.

Abstract

Under noisy conditions, automatic speech recognition (ASR) can greatly benefit from the addition of visual signals coming from a video of the speaker's face. However, when multiple candidate speakers are visible this traditionally requires solving a separate problem, namely active speaker detection (ASD), which entails selecting at each moment in time which of the visible faces corresponds to the audio. Recent work has shown that we can solve both problems simultaneously by employing an attention mechanism over the competing video tracks of the speakers' faces, at the cost of sacrificing some accuracy on active speaker detection. This work closes this gap in active speaker detection accuracy by presenting a single model that can be jointly trained with a multi-task loss. By combining the two tasks during training we reduce the ASD classification accuracy by approximately 25%, while simultaneously improving the ASR performance when compared to the multi-person baseline trained exclusively for ASR.