2000
DOI: 10.1016/s0167-6393(99)00030-8
|View full text |Cite
|
Sign up to set email alerts
|

An efficient search space representation for large vocabulary continuous speech recognition

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
33
0
1

Year Published

2001
2001
2016
2016

Publication Types

Select...
4
3
1

Relationship

1
7

Authors

Journals

citations
Cited by 42 publications
(34 citation statements)
references
References 22 publications
0
33
0
1
Order By: Relevance
“…To decode, a single-pass time synchronous beam search algorithm is used [Demuynck et al, 2000]. For the clean data, the recogniser runs in real time on a 2.0 GHz Pentium 4 processor.…”
Section: Aurora4 Recognisermentioning
confidence: 99%
“…To decode, a single-pass time synchronous beam search algorithm is used [Demuynck et al, 2000]. For the clean data, the recogniser runs in real time on a 2.0 GHz Pentium 4 processor.…”
Section: Aurora4 Recognisermentioning
confidence: 99%
“…In one version our in-house large-vocabulary speakerindependent continuous speech recognizer was used [3,4]. This system is, similar to all modern large vocabulary speech recogniz- ers, based on statistical pattern recognition and integrates all knowledge concerning the spoken input according to the scheme depicted in figure 13.…”
Section: Speech Recognitionmentioning
confidence: 99%
“…Due to the smart representation of the lexicon [3] and the fast evaluation of the acoustic models [4], our recognizer provides realtime decoding in combination with a low latency, even for continuous speech and large vocabularies (20000 or more words). The lowlatency of the recognizer allows for example direct visual feedback to the user by means of frowning eyebrows whenever a language construct that is likely to be misunderstood by the virtual guide, is heard.…”
Section: Speech Recognitionmentioning
confidence: 99%
“…Around 2005 there were several research institutions in the low countries that had developed speech recognition systems for the Dutch language [2,15]. Some were using CGN [20], others used their own databases [10,15].…”
Section: Introductionmentioning
confidence: 99%