Scalable distributed speech recognition using multi-frame GMM-based block quantization.
Kuldip K. PaliwalStephen SoPublished in: INTERSPEECH (2004)
Keyphrases
- multi frame
- scalable distributed
- speaker recognition
- speaker identification
- gaussian mixture model
- motion estimation
- voice activity detection
- point correspondences
- speech recognition
- image sequences
- super resolution
- multiple frames
- speech signal
- optical flow
- computationally efficient
- temporal coherence
- single frame
- motion analysis
- transform coefficients
- file system
- optic flow
- automatic speech recognition
- noisy environments
- mixture model
- filtering algorithm
- background subtraction
- vector quantization
- video frames
- broadcast news
- flow field
- machine learning
- expectation maximization
- state space
- feature vectors
- multimedia