The Development Of The 1994 HTK Large Vocabulary Speech Recognition System

The Development Of The 1994 HTK Large Vocabulary Speech Recognition System

Cached

Download Links

[svr-www.eng.cam.ac.uk]

by P.C. Woodland , C.J. Leggetter , J. J. Odell , V. Valtchev , S.J. Young

Citations:

56 - 5 self

Versions

Version History

Metadata Version 2

User correction supplied by SystemCorrections

Datum	Value	Source
TITLE	The Development Of The 1994 HTK Large Vocabulary Speech Recognition System	user correction - Legacy Corrections
AUTHOR NAME	P.C. Woodland	user correction - Legacy Corrections
AUTHOR AFFIL	Cambridge University Engineering Department,; Trumpington Street, Cambridge, CB2 1PZ	user correction - Legacy Corrections
AUTHOR ADDR	England.	user correction - Legacy Corrections
AUTHOR NAME	C.J. Leggetter	user correction - Legacy Corrections
AUTHOR AFFIL	Cambridge University Engineering Department,; Trumpington Street, Cambridge, CB2 1PZ	user correction - Legacy Corrections
AUTHOR ADDR	England.	user correction - Legacy Corrections
AUTHOR NAME	J. J. Odell	user correction - Legacy Corrections
AUTHOR AFFIL	Cambridge University Engineering Department,; Trumpington Street, Cambridge, CB2 1PZ	user correction - Legacy Corrections
AUTHOR ADDR	England.	user correction - Legacy Corrections
AUTHOR NAME	V. Valtchev	user correction - Legacy Corrections
AUTHOR AFFIL	Cambridge University Engineering Department,; Trumpington Street, Cambridge, CB2 1PZ	user correction - Legacy Corrections
AUTHOR ADDR	England.	user correction - Legacy Corrections
AUTHOR NAME	S.J. Young	user correction - Legacy Corrections
AUTHOR AFFIL	Cambridge University Engineering Department,; Trumpington Street, Cambridge, CB2 1PZ	user correction - Legacy Corrections
AUTHOR ADDR	England.	user correction - Legacy Corrections
ABSTRACT	This paper describes recent developments of the HTK large vocabulary continuous speech recognition system. The system uses tied-state cross-word context-dependent mixture Gaussian HMMs and a dynamic network decoder that can operate in a single pass. In the last year the decoder has been extended to produce word lattices to allow flexible and efficient system development, as well as multi-pass operation for use with computationally expensive acoustic and/or language models. The system vocabulary can now be up to 65k words, the final acoustic models have been extended to be sensitive to more acoustic context (quinphones), a 4-gram language model has been used and unsupervised incremental speaker adaptation incorporated. The resulting system gave the lowest error rates on both the H1-P0 and H1-C1 hub tasks in the November 1994 ARPA CSR evaluation. 1. INTRODUCTION This paper describes recent improvements to the HTK large vocabulary speech recognition system. The system uses state-clustere...	user correction - Legacy Corrections
CITATIONS	7 found	ParsCit 1.0