Title :
The Lincoln robust continuous speech recognizer
Author :
Paul, Douglas B.
Author_Institution :
MIT Lincoln Lab., Lexington, MA, USA
Abstract :
The Lincoln stress-resistant HMM (hidden Markov model) CSR has been extended to large-vocabulary continuous speech for both speaker-dependent (SD) and speaker-independent (SI) tasks. Performance on the DARPA resource management task (991-word vocabulary, perplexity 60 word-pair grammar) is 3.5% word error rate for SD training of word-context-dependent triphone models and 12.6% word error rate for SI training of (word-context-free) tied-mixture triphone models
Keywords :
Markov processes; speech recognition; DARPA resource management task; Lincoln robust continuous speech recognizer; large-vocabulary continuous speech; speaker dependent recognition; speaker independent recognition; speech recognition; stress resistant hidden Markov model; tied-mixture triphone models; word error rate; word-context-dependent triphone models; Databases; Error analysis; Hidden Markov models; Management training; Resource management; Robustness; Speech enhancement; Speech recognition; System testing; Vocabulary;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 1989. ICASSP-89., 1989 International Conference on
Conference_Location :
Glasgow
DOI :
10.1109/ICASSP.1989.266460
