Whisper-Flamingo: Integrating Visual Features into Whisper for Audio-Visual Speech Recognition and Translation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Rouditchenko, Andrew, Gong, Yuan, Thomas, Samuel, Karlinsky, Leonid, Kuehne, Hilde, Feris, Rogerio, Glass, James
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916488373338112
author Rouditchenko, Andrew
Gong, Yuan
Thomas, Samuel
Karlinsky, Leonid
Kuehne, Hilde
Feris, Rogerio
Glass, James
author_facet Rouditchenko, Andrew
Gong, Yuan
Thomas, Samuel
Karlinsky, Leonid
Kuehne, Hilde
Feris, Rogerio
Glass, James
contents Audio-Visual Speech Recognition (AVSR) uses lip-based video to improve performance in noise. Since videos are harder to obtain than audio, the video training data of AVSR models is usually limited to a few thousand hours. In contrast, speech models such as Whisper are trained with hundreds of thousands of hours of data, and thus learn a better speech-to-text decoder. The huge training data difference motivates us to adapt Whisper to handle video inputs. Inspired by Flamingo which injects visual features into language models, we propose Whisper-Flamingo which integrates visual features into the Whisper speech recognition and translation model with gated cross attention. Our models achieve state-of-the-art ASR WER (0.68%) and AVSR WER (0.76%) on LRS3, and state-of-the-art ASR WER (1.3%) and AVSR WER (1.4%) on LRS2. Audio-visual Whisper-Flamingo outperforms audio-only Whisper on English speech recognition and En-X translation for 6 languages in noisy conditions. Moreover, Whisper-Flamingo is versatile and conducts all of these tasks using one set of parameters, while prior methods are trained separately on each language.
format Preprint
id arxiv_https___arxiv_org_abs_2406_10082
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Whisper-Flamingo: Integrating Visual Features into Whisper for Audio-Visual Speech Recognition and Translation
Rouditchenko, Andrew
Gong, Yuan
Thomas, Samuel
Karlinsky, Leonid
Kuehne, Hilde
Feris, Rogerio
Glass, James
Audio and Speech Processing
Computer Vision and Pattern Recognition
Sound
Audio-Visual Speech Recognition (AVSR) uses lip-based video to improve performance in noise. Since videos are harder to obtain than audio, the video training data of AVSR models is usually limited to a few thousand hours. In contrast, speech models such as Whisper are trained with hundreds of thousands of hours of data, and thus learn a better speech-to-text decoder. The huge training data difference motivates us to adapt Whisper to handle video inputs. Inspired by Flamingo which injects visual features into language models, we propose Whisper-Flamingo which integrates visual features into the Whisper speech recognition and translation model with gated cross attention. Our models achieve state-of-the-art ASR WER (0.68%) and AVSR WER (0.76%) on LRS3, and state-of-the-art ASR WER (1.3%) and AVSR WER (1.4%) on LRS2. Audio-visual Whisper-Flamingo outperforms audio-only Whisper on English speech recognition and En-X translation for 6 languages in noisy conditions. Moreover, Whisper-Flamingo is versatile and conducts all of these tasks using one set of parameters, while prior methods are trained separately on each language.
title Whisper-Flamingo: Integrating Visual Features into Whisper for Audio-Visual Speech Recognition and Translation
topic Audio and Speech Processing
Computer Vision and Pattern Recognition
Sound
url https://arxiv.org/abs/2406.10082