Hervé Bourlard
Automatic speech recognition bases its models on the acoustic features derived from the speech signal. Some have investigated replacing or supplementing these features with information that can not be precisely measured (articulator positions, pitch, gende ...
IDIAP2001