C82-1035 constraints are frequently used to help acoustic recognition by reducing the number of possibilities
W99-0617 single tokens in order to improve acoustic recognition of them . The result of the first
J10-4002 pronounced with little effort and their acoustic recognition is less reliable . In stressed
H89-1006 improved methods and models for acoustic recognition of continuous speech . Most of
H90-1076 or epenthetic stops . Guided by acoustic recognition 2 errors , we have devised a
J88-2015 combination of the optical and acoustic recognition systems could result in 95 %
J87-1020 indeterminacies and inaccuracies of acoustic recognition must be handled in an integral
C86-1138 indeterminacies and inaccuracies of acoustic recognition must be handled in an integral
J92-3011 predicted phone ) . In this way acoustic recognition can be resumed . The algorithm
H89-2027 a NL system by any reasonable acoustic recognition component . Thus , much of the
H94-1088 models for speaker-independent acoustic recognition of spontaneously-produced , continuous
H93-1079 models for speaker-independent acoustic recognition of spontaneously-produced , continuous
H91-1080 imProved methods and models for acoustic recognition of continuous speech . The work
H89-2027 generate N-best lists from a real acoustic recognition system , then we can ask the
H92-1097 models for speaker-independent acoustic recognition of spontaneously-produced , :
H89-2027 further processing required from the acoustic recognition component , the interface between
H92-1010 between the language model and the acoustic recognition was made , and resulted in a
H89-2058 improved methods and models for acoustic recognition of continuous speech . The work
H90-1079 improved methods and models for acoustic recognition of continuous speech . The work
hide detail