Low memory acoustic models for HMM based speech recognition.
Tommi LahtiOlli ViikkiMarcel VasilachePublished in: INTERSPEECH (2003)
Keyphrases
- low memory
- acoustic models
- speech recognition
- hidden markov models
- speech recognizer
- speaker independent
- automatic speech recognition
- memory requirements
- speech signal
- pattern recognition
- discriminative training
- language model
- neural network
- noisy environments
- speech recognition systems
- mixture model
- visual features
- em algorithm
- speaker identification
- image classification
- denoising