Multilingual acoustic modeling for speech recognition based on subspace Gaussian Mixture Models.
Lukás BurgetPetr SchwarzMohit AgarwalPinar AkyaziKai FengArnab GhoshalOndrej GlembekNagendra K. GoelMartin KarafiátDaniel PoveyAriya RastrowRichard C. RoseSamuel ThomasPublished in: ICASSP (2010)
Keyphrases
- speech recognition
- gaussian mixture model
- speaker identification
- speaker recognition
- speaker independent
- mel frequency cepstral coefficients
- speech recognition systems
- covariance matrices
- mixture model
- speech recognizers
- language model
- feature space
- hidden markov models
- speech signal
- pattern recognition
- speech synthesis
- em algorithm
- automatic speech recognition
- acoustic models
- feature vectors
- noisy environments
- speech recognizer
- maximum likelihood
- probabilistic neural network
- expectation maximization
- gaussian distribution
- noisy speech
- principal component analysis
- feature extraction
- dimensionality reduction
- speaker diarization
- speaker verification
- language identification
- acoustic features
- low dimensional
- speech retrieval