Advances in Online Audio-Visual Meeting Transcription
Abstract
This paper describes a system that generates speaker-annotated transcripts of meetings by using a microphone array and a 360-degree camera. The hallmark of the system is its ability to handle overlapped speech, which has been an unsolved problem in realistic settings for over a decade. We show that this problem can be addressed by using a continuous speech separation approach. In addition, we describe an online audio-visual speaker diarization method that leverages face tracking and identification, sound source localization, speaker identification, and, if available, prior speaker information for robustness to various real world challenges. All components are integrated in a meeting transcription framework called SRD, which stands for "separate, recognize, and diarize". Experimental results using recordings of natural meetings involving up to 11 attendees are reported. The continuous speech separation improves a word error rate (WER) by 16.1% compared with a highly tuned beamformer. When a complete list of meeting attendees is available, the discrepancy between WER and speaker-attributed WER is only 1.0%, indicating accurate word-to-speaker association. This increases marginally to 1.6% when 50% of the attendees are unknown to the system.
Keywords
Cite
@article{arxiv.1912.04979,
title = {Advances in Online Audio-Visual Meeting Transcription},
author = {Takuya Yoshioka and Igor Abramovski and Cem Aksoylar and Zhuo Chen and Moshe David and Dimitrios Dimitriadis and Yifan Gong and Ilya Gurvich and Xuedong Huang and Yan Huang and Aviv Hurvitz and Li Jiang and Sharon Koubi and Eyal Krupka and Ido Leichter and Changliang Liu and Partha Parthasarathy and Alon Vinnikov and Lingfeng Wu and Xiong Xiao and Wayne Xiong and Huaming Wang and Zhenghao Wang and Jun Zhang and Yong Zhao and Tianyan Zhou},
journal= {arXiv preprint arXiv:1912.04979},
year = {2019}
}
Comments
To appear in Proc. IEEE ASRU Workshop 2019