Retrieving radio news broadcasts in Danish: accuracy and categorization of unrecognized words

Morten Hertzum; Haakon Lund; Rasmus Troelsgård

doi:10.1145/3010915.3010972

Retrieving radio news broadcasts in Danish: accuracy and categorization of unrecognized words

Morten Hertzum, Haakon Lund, Rasmus Troelsgård

Institut for Kommunikation - Forskning

Abstract

Digital archives of radio news broadcasts can possibly be made searchable by combining speech recognition with information retrieval. We explore this possibility for the retrieval of news broadcasts in Danish. An average of 84% of the words in the broadcasts was recognized. Most of the unrecognized words were compounds, names, and other words that appear of value to retrieval. Thus, the set of words describing a broadcast has to be expanded to compensate for the recognition errors. We discuss doing this by exploiting the alternative matches from the speech recognizer and by extracting words from a related corpus

Originalsprog	Engelsk
Titel	OzCHI'16 : The 28th Australian Conference on Compute-Human Interaction
Antal sider	5
Udgivelsessted	New York
Forlag	ACM
Publikationsdato	29 nov. 2016
Sider	160-164
ISBN (Elektronisk)	978-1-4503-4618-4
DOI	https://doi.org/10.1145/3010915.3010972
Status	Udgivet - 29 nov. 2016
Begivenhed	Australian Conference on Human-Computer Interaction - Launceston, Australien Varighed: 29 nov. 2016 → 2 dec. 2016 Konferencens nummer: 28 http://www.ozchi.org/2016/index.html

Konference

Konference	Australian Conference on Human-Computer Interaction
Nummer	28
Lokation	Launceston
Land/Område	Australien
Periode	29/11/2016 → 02/12/2016
Internetadresse	http://www.ozchi.org/2016/index.html

Adgang til dokumentet

10.1145/3010915.3010972

Australian Conference on Human-Computer Interaction
Haakon Lund (Taler)
30 nov. 2016
Aktivitet: Deltagelse i eller arrangement af en begivenhed - typer › Deltagelse i workshop, seminar og kursus

Citationsformater

@inproceedings{d51a1daf83f7498aa5bb34e2dd2443c6,

title = "Retrieving radio news broadcasts in Danish: accuracy and categorization of unrecognized words",

abstract = "Digital archives of radio news broadcasts can possibly be made searchable by combining speech recognition with information retrieval. We explore this possibility for the retrieval of news broadcasts in Danish. An average of 84% of the words in the broadcasts was recognized. Most of the unrecognized words were compounds, names, and other words that appear of value to retrieval. Thus, the set of words describing a broadcast has to be expanded to compensate for the recognition errors. We discuss doing this by exploiting the alternative matches from the speech recognizer and by extracting words from a related corpus",

author = "Morten Hertzum and Haakon Lund and Rasmus Troelsg{\aa}rd",

year = "2016",

month = nov,

day = "29",

doi = "10.1145/3010915.3010972",

language = "English",

pages = "160--164",

booktitle = "OzCHI'16",

publisher = "ACM",

note = "Australian Conference on Human-Computer Interaction , OzCHI 2016 ; Conference date: 29-11-2016 Through 02-12-2016",

url = "http://www.ozchi.org/2016/index.html",

}

TY - GEN

T1 - Retrieving radio news broadcasts in Danish

T2 - Australian Conference on Human-Computer Interaction

AU - Hertzum, Morten

AU - Lund, Haakon

AU - Troelsgård, Rasmus

N1 - Conference code: 28

PY - 2016/11/29

Y1 - 2016/11/29

N2 - Digital archives of radio news broadcasts can possibly be made searchable by combining speech recognition with information retrieval. We explore this possibility for the retrieval of news broadcasts in Danish. An average of 84% of the words in the broadcasts was recognized. Most of the unrecognized words were compounds, names, and other words that appear of value to retrieval. Thus, the set of words describing a broadcast has to be expanded to compensate for the recognition errors. We discuss doing this by exploiting the alternative matches from the speech recognizer and by extracting words from a related corpus

AB - Digital archives of radio news broadcasts can possibly be made searchable by combining speech recognition with information retrieval. We explore this possibility for the retrieval of news broadcasts in Danish. An average of 84% of the words in the broadcasts was recognized. Most of the unrecognized words were compounds, names, and other words that appear of value to retrieval. Thus, the set of words describing a broadcast has to be expanded to compensate for the recognition errors. We discuss doing this by exploiting the alternative matches from the speech recognizer and by extracting words from a related corpus

U2 - 10.1145/3010915.3010972

DO - 10.1145/3010915.3010972

M3 - Article in proceedings

SP - 160

EP - 164

BT - OzCHI'16

PB - ACM

CY - New York

Y2 - 29 November 2016 through 2 December 2016

ER -

Retrieving radio news broadcasts in Danish: accuracy and categorization of unrecognized words

Abstract

Konference

Adgang til dokumentet

Fingeraftryk

Aktiviteter

Australian Conference on Human-Computer Interaction

Citationsformater