Audio information access from meeting rooms

Steve Renals; Dan  Ellis

doi:10.1109/ICASSP.2003.1202750

Audio information access from meeting rooms

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

We investigate approaches to accessing information from the streams of audio data that result from multi-channel recordings of meetings. The methods investigated use word-level transcriptions, and information derived from models of speaker activity and speaker turn patterns. Our experiments include spoken document retrieval for meetings, automatic structuring of meetings based on self-similarity matrices of speaker turn patterns and a simple model of speaker activity. Meeting recordings are rich in both lexical and non-lexical information; our results illustrate some novel kinds of analysis made possible by a transcribed corpus of natural meetings.

Original language	English
Title of host publication	2003 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings of the
Subtitle of host publication	ICASSP'03
Publisher	Institute of Electrical and Electronics Engineers
Pages	744-747
Volume	4
ISBN (Print)	0-7803-7663-3
DOIs	https://doi.org/10.1109/ICASSP.2003.1202750
Publication status	Published - 2003
Event	2003 IEEE International Conference on Acoustics, Speech, and Signal Processing - Hong Kong Exhibition and Convention Centre, Hong Kong, Hong Kong Duration: 6 Apr 2003 → 10 Apr 2003

Conference

Conference	2003 IEEE International Conference on Acoustics, Speech, and Signal Processing
Country/Territory	Hong Kong
City	Hong Kong
Period	6/04/03 → 10/04/03

Access to Document

10.1109/ICASSP.2003.1202750

icassp03-mtgSubmitted manuscript, 73.5 KB

http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=1202750

Cite this

@inproceedings{d367a2333c3949a48235f77fc740f049,

title = "Audio information access from meeting rooms",

abstract = "We investigate approaches to accessing information from the streams of audio data that result from multi-channel recordings of meetings. The methods investigated use word-level transcriptions, and information derived from models of speaker activity and speaker turn patterns. Our experiments include spoken document retrieval for meetings, automatic structuring of meetings based on self-similarity matrices of speaker turn patterns and a simple model of speaker activity. Meeting recordings are rich in both lexical and non-lexical information; our results illustrate some novel kinds of analysis made possible by a transcribed corpus of natural meetings.",

author = "Steve Renals and Dan Ellis",

year = "2003",

doi = "10.1109/ICASSP.2003.1202750",

language = "English",

isbn = "0-7803-7663-3",

volume = "4",

pages = "744--747",

booktitle = "2003 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings of the",

publisher = "Institute of Electrical and Electronics Engineers",

address = "United States",

note = "2003 IEEE International Conference on Acoustics, Speech, and Signal Processing ; Conference date: 06-04-2003 Through 10-04-2003",

}

Renals, S & Ellis, D 2003, Audio information access from meeting rooms. in 2003 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings of the: ICASSP'03. vol. 4, Institute of Electrical and Electronics Engineers, pp. 744-747, 2003 IEEE International Conference on Acoustics, Speech, and Signal Processing, Hong Kong, Hong Kong, 6/04/03. https://doi.org/10.1109/ICASSP.2003.1202750

TY - GEN

T1 - Audio information access from meeting rooms

AU - Renals, Steve

AU - Ellis, Dan

PY - 2003

Y1 - 2003

N2 - We investigate approaches to accessing information from the streams of audio data that result from multi-channel recordings of meetings. The methods investigated use word-level transcriptions, and information derived from models of speaker activity and speaker turn patterns. Our experiments include spoken document retrieval for meetings, automatic structuring of meetings based on self-similarity matrices of speaker turn patterns and a simple model of speaker activity. Meeting recordings are rich in both lexical and non-lexical information; our results illustrate some novel kinds of analysis made possible by a transcribed corpus of natural meetings.

AB - We investigate approaches to accessing information from the streams of audio data that result from multi-channel recordings of meetings. The methods investigated use word-level transcriptions, and information derived from models of speaker activity and speaker turn patterns. Our experiments include spoken document retrieval for meetings, automatic structuring of meetings based on self-similarity matrices of speaker turn patterns and a simple model of speaker activity. Meeting recordings are rich in both lexical and non-lexical information; our results illustrate some novel kinds of analysis made possible by a transcribed corpus of natural meetings.

U2 - 10.1109/ICASSP.2003.1202750

DO - 10.1109/ICASSP.2003.1202750

M3 - Conference contribution

SN - 0-7803-7663-3

VL - 4

SP - 744

EP - 747

BT - 2003 IEEE International Conference on Acoustics, Speech, and Signal Processing. Proceedings of the

PB - Institute of Electrical and Electronics Engineers

T2 - 2003 IEEE International Conference on Acoustics, Speech, and Signal Processing

Y2 - 6 April 2003 through 10 April 2003

ER -

Audio information access from meeting rooms

Abstract

Conference

Access to Document

Fingerprint

Cite this