Speaker segmentation based on Bayesian Information Criterion (BIC) for Slovenian BroadCast News transcription system /
The article presents initial speaker segmentation experiments on the Slovenian Broadcast News Database. Speaker segmentation was performed with the goal of segmenting audio signals into speaker utterance groups or clusters which are used for speaker adaptation of continuous speech recognition system...
Shranjeno v:
| Main Authors: | , |
|---|---|
| Format: | Book Chapter |
| Jezik: | English |
| Teme: | |
| Sorodne knjige/članki: | Vsebovano v:
Advances in speech technology |
| Oznake: |
Označite
Brez oznak, prvi označite!
|
| LEADER | 01768naa a2200205 ib4500 | ||
|---|---|---|---|
| 001 | 12464918 | ||
| 003 | SI-MaCOB | ||
| 005 | 20080820000000.0 | ||
| 008 | 080820s2008 xv |||||||||||||| ||eng c | ||
| 040 | |a KTFMB |b slv |c SI-MaIIZ |d KTFMB |e ppiak | ||
| 080 | |a 004.93 | ||
| 100 | 1 | |a Grašič, Matej, |d 1981- |4 aut | |
| 245 | 0 | 0 | |a Speaker segmentation based on Bayesian Information Criterion (BIC) for Slovenian BroadCast News transcription system / |c Matej Grašič and Marko Kos. |
| 300 | |a Str. 53-59. | ||
| 504 | |a Bibliografija: str. 59. | ||
| 520 | |a The article presents initial speaker segmentation experiments on the Slovenian Broadcast News Database. Speaker segmentation was performed with the goal of segmenting audio signals into speaker utterance groups or clusters which are used for speaker adaptation of continuous speech recognition systems. Initial experiments feature only speaker turn segmentation using the Bayesian Information Criteria. Evaluation was performed using the ELIS evaluation tool on a subset of the Slovenian BNSI Broadcast News speech database. The accuracy results are promising and comparable with other transcription systems. For online operations computational speed improvements are needed to achieve real-time performance. | ||
| 653 | 0 | |a avtomatsko razpoznavanje govora |a jezikovne baze podatkov |a govor |a segmentacija govora |a audio signali | |
| 653 | 0 | |a automatic speech recognition |a speech segmentation |a speaker segmentation |a broadcast news database |a Bayesian information criterion | |
| 700 | 1 | 2 | |a Kos, Marko, |d 1981- |4 aut |
| 773 | 0 | |a International Workshop Advances in Speech Technology (14 : 2007 : Maribor) |t Advances in speech technology |d Maribor : Faculty of Electrical Engineering and Computer Science, 2008 |w 61075201 |z 9789612481056 |g Str. 53-59 | |