Distributed speech recognition using noise-robust MFCC and traps-estimated manner features.
Pratibha JainHynek HermanskyBrian KingsburyPublished in: INTERSPEECH (2002)
Keyphrases
- speech recognition
- noisy environments
- feature extraction
- feature set
- speech signal
- feature vectors
- intensity gradient
- mel frequency cepstral coefficients
- estimation error
- speech recognition systems
- image noise
- low level
- noisy data
- edge detection
- missing data
- cooperative
- speaker identification
- multi agent
- noise level
- image processing
- image features
- classification accuracy
- registration errors
- linear predictive
- cepstral coefficients
- speech enhancement
- speech recognizer
- speaker verification
- extracting features
- audio features
- audio visual
- noise reduction
- gaussian mixture model
- keypoints