|
Volumn , Issue , 2003, Pages 573-576
|
A trainable speech enhancement technique based on mixture models for speech and noise
|
Author keywords
[No Author keywords available]
|
Indexed keywords
ADAPTIVE SYSTEMS;
BAYESIAN NETWORKS;
IMAGE SEGMENTATION;
INFERENCE ENGINES;
MAXIMUM PRINCIPLE;
MEAN SQUARE ERROR;
MIXTURES;
NOISE POLLUTION;
SIGNAL RECEIVERS;
SIGNAL TO NOISE RATIO;
SPEECH;
SPEECH ENHANCEMENT;
CLOSED FORM SOLUTIONS;
EXPECTATION-MAXIMIZATION ALGORITHMS;
MINIMUM MEAN SQUARE ERROR ESTIMATIONS;
NONSTATIONARY NOISE;
OPERATIONAL ENVIRONMENTS;
SPECTRAL COEFFICIENTS;
SPECTRAL MAGNITUDES;
TIME FREQUENCY CHARACTERISTICS;
SPEECH COMMUNICATION;
|
EID: 85009165868
PISSN: None
EISSN: None
Source Type: Conference Proceeding
DOI: None Document Type: Conference Paper |
Times cited : (4)
|
References (8)
|