fMLLR based feature-space speaker adaptation of DNN acoustic models.
Sree Hari Krishnan ParthasarathiBjörn HoffmeisterSpyros MatsoukasArindam MandalNikko StromSri GarimellaPublished in: INTERSPEECH (2015)
Keyphrases
- acoustic models
- speaker adaptation
- speech recognition
- feature space
- speaker independent
- speech recognizer
- automatic speech recognition
- hidden markov models
- speaker identification
- language model
- feature vectors
- gaussian mixture model
- feature extraction
- spoken language
- principal component analysis
- feature selection
- digit recognition
- pattern recognition
- broadcast news
- speech signal
- training process
- speech recognition systems
- image retrieval
- training set
- machine learning
- discriminative training
- maximum likelihood
- support vector machine
- information retrieval