https://scholars.lib.ntu.edu.tw/handle/123456789/498618
標題: | Rapid speaker adaptation using a priori knowledge by eigenspace analysis of MLLR parameters | 作者: | Wang, N.J.-C. Lee, S.S.-M. Seide, F. LIN-SHAN LEE |
公開日期: | 2001 | 卷: | 1 | 起(迄)頁: | 345-348 | 來源出版物: | ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings | 摘要: | This paper considers the problem of rapid speaker adaptation in speech recognition. In particular, we exploit an approach based on combination of transformations, which utilizes the concepts of both maximum likelihood linear regression (MLLR) and eigenvoice adaptation. We analyze three different possible methods to realize the concept, and formulate a fast algorithm of maximum likelihood coefficient estimation for test speakers. It was found that the best approach can properly utilize the a priori knowledge of speaker-independent models in constructing the eigenspace for speaker characteristics, while using MLLR matrices in representing the specific speakers so as to reduce the on-line memory and computation requirement of the adaptation phase. This best approach leads to identical models as eigenvoice adaptation that is based on MLLR-adapted speaker models. The experimental results and discussions also provide a good analysis towards integration of MLLR and eigenvoice approaches. |
URI: | https://scholars.lib.ntu.edu.tw/handle/123456789/498618 https://www.scopus.com/inward/record.uri?eid=2-s2.0-0034843060&doi=10.1109%2fICASSP.2001.940838&partnerID=40&md5=739af01ee4b88914aebabaa1998e69b1 |
ISSN: | 15206149 | DOI: | 10.1109/ICASSP.2001.940838 | SDG/關鍵字: | A priori knowledge; Eigenspace analysis; Eigenvoice adaptation; Maximum likelihood linear regression; Rapid speaker adaptation; Algorithms; Eigenvalues and eigenfunctions; Mathematical models; Maximum likelihood estimation; Speech analysis; Speech recognition |
顯示於: | 電機工程學系 |
在 IR 系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。