Title :
Efficient Generation of high-order context-dependent Weighted Finite State Transducers for Speech Recognition
Author :
Schuster, Mike ; Hori, Takaaki
Author_Institution :
NTT Commun. Sci. Labs., NTT Corp., Kyoto, Japan
fDate :
March 18-23, 2005
Keywords :
decision trees; finite state machines; speech recognition; 125 MB; context-dependent acoustic models; full cross-word quinphones; high-order context dependent weighted finite state transducers; high-order full cross-word models; one-pass time-synchronous search; phonetic decision trees; real-time recognition; recognition network size; recognition rate; speech recognition; tied-state triphone models; Acoustic testing; Context modeling; Databases; Decision trees; Decoding; Frequency; Laboratories; Memory management; Speech recognition; Transducers;
Conference_Titel :
Acoustics, Speech, and Signal Processing, 2005. Proceedings. (ICASSP '05). IEEE International Conference on
Print_ISBN :
0-7803-8874-7
DOI :
10.1109/ICASSP.2005.1415085
