Dealing with acoustic mismatch for training multilingual subspace Gaussian mixture models for speech recognition.
Aanchan MohanSina Hamidi GhalehjeghRichard C. RosePublished in: ICASSP (2012)
Keyphrases
- speech recognition
- gaussian mixture model
- acoustic models
- speaker identification
- speaker recognition
- mel frequency cepstral coefficients
- speaker independent
- speech recognition systems
- mixture model
- speech signal
- feature space
- language model
- speech recognizers
- hidden markov models
- automatic speech recognition
- isolated word
- pattern recognition
- speech recognizer
- em algorithm
- maximum likelihood
- principal component analysis
- probabilistic neural network
- feature vectors
- acoustic features
- noisy environments
- speech synthesis
- noisy speech
- low dimensional
- gaussian distribution
- training set
- expectation maximization
- audio signal
- feature extraction
- machine learning
- probabilistic model
- classification accuracy
- supervised learning