Abstract
This paper proposes an unsupervised speaker clustering system for duration of speech signals below 4 seconds. For determining whether two collected speech sections uttered from the same speaker or not, our previous SVM training miss-classification rate (STMR) is adopted to evaluate the data separability between two different speakers. This paper also proposes a hierarchical extract and merge (HEM) clustering method to reduce agglomeration time and enhance the clustering purity. Experiment results show the average speaker purity (ASP) and average cluster purity (ACP) are both better than the CE manner with the GMM training miss-classification rates (GTMR) for 2 to 4 seconds short speech sections.
| Original language | English |
|---|---|
| Title of host publication | Proceedings - 4th International Conference on Genetic and Evolutionary Computing, ICGEC 2010 |
| Pages | 606-609 |
| Number of pages | 4 |
| DOIs | |
| Publication status | Published - 2010 Dec 1 |
| Event | 4th International Conference on Genetic and Evolutionary Computing, ICGEC 2010 - Shenzhen, China Duration: 2010 Dec 13 → 2010 Dec 15 |
Other
| Other | 4th International Conference on Genetic and Evolutionary Computing, ICGEC 2010 |
|---|---|
| Country | China |
| City | Shenzhen |
| Period | 10-12-13 → 10-12-15 |
Fingerprint
All Science Journal Classification (ASJC) codes
- Computational Theory and Mathematics
- Theoretical Computer Science
Cite this
}
Unsupervised speaker clustering using SVM training missclassification rate for meeting short-duration speech signals. / Lin, Po Chuan; Jui, Yeh Yi; Ying, Tsai Sung; Chen, Yeong Chin; Wu, Menq-Jiun.
Proceedings - 4th International Conference on Genetic and Evolutionary Computing, ICGEC 2010. 2010. p. 606-609 5715505.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
TY - GEN
T1 - Unsupervised speaker clustering using SVM training missclassification rate for meeting short-duration speech signals
AU - Lin, Po Chuan
AU - Jui, Yeh Yi
AU - Ying, Tsai Sung
AU - Chen, Yeong Chin
AU - Wu, Menq-Jiun
PY - 2010/12/1
Y1 - 2010/12/1
N2 - This paper proposes an unsupervised speaker clustering system for duration of speech signals below 4 seconds. For determining whether two collected speech sections uttered from the same speaker or not, our previous SVM training miss-classification rate (STMR) is adopted to evaluate the data separability between two different speakers. This paper also proposes a hierarchical extract and merge (HEM) clustering method to reduce agglomeration time and enhance the clustering purity. Experiment results show the average speaker purity (ASP) and average cluster purity (ACP) are both better than the CE manner with the GMM training miss-classification rates (GTMR) for 2 to 4 seconds short speech sections.
AB - This paper proposes an unsupervised speaker clustering system for duration of speech signals below 4 seconds. For determining whether two collected speech sections uttered from the same speaker or not, our previous SVM training miss-classification rate (STMR) is adopted to evaluate the data separability between two different speakers. This paper also proposes a hierarchical extract and merge (HEM) clustering method to reduce agglomeration time and enhance the clustering purity. Experiment results show the average speaker purity (ASP) and average cluster purity (ACP) are both better than the CE manner with the GMM training miss-classification rates (GTMR) for 2 to 4 seconds short speech sections.
UR - http://www.scopus.com/inward/record.url?scp=79952571088&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=79952571088&partnerID=8YFLogxK
U2 - 10.1109/ICGEC.2010.155
DO - 10.1109/ICGEC.2010.155
M3 - Conference contribution
SN - 9780769542812
SP - 606
EP - 609
BT - Proceedings - 4th International Conference on Genetic and Evolutionary Computing, ICGEC 2010
ER -