Feature compensation employing online GMM adaptation for speech recognition in unknown severely adverse environments.
Wooil KimJohn H. L. HansenPublished in: ICASSP (2012)
Keyphrases
- speech recognition
- speaker identification
- speaker adaptation
- speaker recognition
- hidden markov models
- gaussian mixture model
- speech recognition technology
- cepstral coefficients
- speech signal
- speech recognizer
- feature vectors
- pattern recognition
- automatic speech recognition
- language model
- voice activity detection
- speech understanding
- speech synthesis
- speech processing
- noisy environments
- keyword spotting
- speech recognition systems
- speaker independent
- mel frequency cepstral coefficients
- neural network
- speaker dependent
- speech retrieval
- maximum likelihood
- speech recognizers
- probabilistic neural network
- speech recognition errors
- information retrieval