Multi-Frame GMM-Based Block Quantisation of Line Spectral Frequencies for Wideband Speech Coding.
Stephen SoKuldip K. PaliwalPublished in: ICASSP (1) (2005)
Keyphrases
- multi frame
- speech signal
- linear prediction
- speaker identification
- linear predictive
- cepstral coefficients
- speaker recognition
- speech recognition
- motion estimation
- mel frequency cepstral coefficients
- spectral analysis
- linear predictive coding
- point correspondences
- gaussian mixture model
- super resolution
- image sequences
- automatic speech recognition
- spectral features
- multiple frames
- computationally efficient
- motion analysis
- temporal coherence
- speech quality
- hidden markov models
- single frame
- planar scenes
- optic flow
- filtering algorithm
- optical flow
- voice activity detection
- computer vision
- mixture model
- acoustic features
- video frames
- subband
- motion field
- signal to noise ratio
- background subtraction
- three dimensional
- image processing