Building an effective corpus by using acoustic space visualization (COSMOS) method [speech recognition applications]
作者： G. Nagino , M. Shozakai
DOI: 10.1109/ICASSP.2005.1415147
关键词: Acoustic space 、 Speech recognition 、 Loudspeaker 、 Hidden Markov model 、 Cosmos 、 Data visualization 、 Visualization 、 Space (punctuation) 、 Space technology 、 Computer science 、 Acoustic model
摘要: This paper proposes the technique of building an effective corpus with lower cost by using method visualizing multiple HMM acoustic models into a 2D space ("COSMOS" method: comprehensive map objective signal, previously sound) method. In experiment this paper, adapted 533 male speakers are made small quantity voice samples (10 words) per speaker. Then plotted (called COSMOS map) featuring total is generated utilizing A was built selecting 200 located only in periphery distribution and collecting (165 The model trained from showed higher performance than one other selected randomly or all map.
sci-hub.se 参考文章(6)
Makoto Shozakai, Goshu Nagino, Design of ready-made acoustic model library by two-dimensional visualization of acoustic space. conference of the international speech communication association. ,(2004)
A.K. Jain, P.W. Duin, Jianchang Mao, Statistical pattern recognition: a review IEEE Transactions on Pattern Analysis and Machine Intelligence. ,vol. 22, pp. 4- 37 ,(2000) , 10.1109/34.824819
J.W. Sammon, A Nonlinear Mapping for Data Structure Analysis IEEE Transactions on Computers. ,vol. 18, pp. 401- 409 ,(1969) , 10.1109/T-C.1969.222678
Christopher J Leggetter, Philip C Woodland, None, Maximum likelihood linear regression for speaker adaptation of continuous density hidden Markov models Computer Speech & Language. ,vol. 9, pp. 171- 185 ,(1995) , 10.1006/CSLA.1995.0010
Lou Boves, Herman J. M. Steeneken, Arkadiusz Nagórski, Optimal selection of speech data for Automatic Speech Recognition systems international conference on spoken language processing. pp. 2473- 2476 ,(2002)
MaoJianchang, K JainAnil, P W DuinRobert, Statistical Pattern Recognition IEEE Transactions on Pattern Analysis and Machine Intelligence. ,(2000) , 10.1109/34.824819