Abstract | ||
---|---|---|
We have already proposed the application of tree-structured speaker clustering to supervised speaker adaptation. This paper proposes its application to unsupervised speaker adaptation and speaker-independent (SI) speech recognition. This clustering involves the selection of a speaker cluster from among multiple reference speaker clusters arranged in a tree structure. Cluster selection, unlike parameter training, enables quick adaptation using only a small amount of training data. This method was applied to a hidden Markov network (HMnet) and evaluated in Japanese phoneme and phrase recognition experiments. Results show effective unsupervised speaker adaptation using only 5 s calibration speech. In the SI speech recognition experiments, the method reduced the error rate by 8.5% compared with the conventional speaker-independent speech recognition method. (C) 1996 Academic Press Limited. |
Year | DOI | Venue |
---|---|---|
1996 | 10.1006/csla.1996.0004 | COMPUTER SPEECH AND LANGUAGE |
Field | DocType | Volume |
Training set,Computer science,Computational linguistics,Word error rate,Phrase,Speech recognition,Tree structure,Hidden Markov model,Cluster analysis,Speaker adaptation | Journal | 10 |
Issue | ISSN | Citations |
1 | 0885-2308 | 19 |
PageRank | References | Authors |
1.19 | 0 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Tetsuo Kosaka | 1 | 80 | 16.41 |
Shoichi Matsunaga | 2 | 164 | 36.02 |
Shigeki Sagayama | 3 | 1217 | 137.97 |