Subspace Gaussian Mixture Models for speech recognition.
Daniel PoveyLukás BurgetMohit AgarwalPinar AkyaziKai FengArnab GhoshalOndrej GlembekNagendra K. GoelMartin KarafiátAriya RastrowRichard C. RosePetr SchwarzSamuel ThomasPublished in: ICASSP (2010)
Keyphrases
- speech recognition
- gaussian mixture model
- covariance matrices
- feature space
- speaker identification
- speaker recognition
- mixture model
- language model
- hidden markov models
- pattern recognition
- speech synthesis
- feature vectors
- automatic speech recognition
- speech signal
- em algorithm
- maximum likelihood
- expectation maximization
- speech recognizer
- principal component analysis
- dimensionality reduction
- low dimensional
- probability density function
- noisy environments
- feature extraction
- speech recognition systems
- computer vision
- neural network
- gaussian distribution
- high dimensional
- image processing
- mel frequency cepstral coefficients
- speaker independent
- speaker adaptation