Hier finden Sie wissenschaftliche Publikationen aus den Fraunhofer-Instituten.

Contextual verification for open vocabulary spoken term detection

: Schneider, D.; Mertens, T.; Larson, M.; Köhler, J.

Fulltext urn:nbn:de:0011-n-1474699 (345 KByte PDF)
MD5 Fingerprint: b2572b0a1ee33751aa31a1f539866209
Created on: 9.12.2010

International Speech Communication Association -ISCA-:
11th Annual Conference of the International Speech Communication Association, Interspeech 2010. Proceedings. Vol.1 : 26 - 30 September, 2010, Makuhari, Chiba, Japan
Red Hook, NY: Curran, 2011
ISBN: 978-1-617-82123-3
International Speech Communication Association (Annual Conference) <11, 2010, Makuhari>
Conference Paper, Electronic Publication
Fraunhofer IAIS ()
Spracherkennung; speech recognition; Audiomining; spoken term detection; OOV; subword ASR

In spoken term detection, subword speech recognition is a viable means for addressing the out-of-vocabulary (OOV) problem at query time. Applying fuzzy error compensation techniques is needed for coping with inevitable recognition errors, but can lead to high false alarm rates especially for short queries. We propose two novel methods which reject false alarms based on the context of the hypothesized result and the distance to phonetically similar queries. Using the proposed methods, we obtain an increase in precision of 11% absolute at equal recall.