Rapid feature space MLLR speaker adaptation for deep neural network acoustic modeling.
Shilei ZhangYong QinPublished in: ICPR (2016)
Keyphrases
- speaker adaptation
- speech recognition
- speaker independent
- neural network
- feature space
- maximum likelihood
- automatic speech recognition
- pattern recognition
- speech recognizer
- hidden markov models
- speech recognition systems
- speaker identification
- feature selection
- speaker dependent
- principal component analysis
- speech synthesis
- computer vision
- acoustic features
- classification accuracy
- high dimensional
- gaussian mixture model
- image representation
- visual features
- text classification
- image classification