Maximum likelihood noise HMMm estimation in model-based robust speech recognition.
Martin GraciarenaPublished in: INTERSPEECH (2000)
Keyphrases
- speech recognition
- noisy environments
- maximum likelihood
- speech signal
- additive noise
- automatic speech recognition
- speaker adaptation
- estimation error
- hidden markov models
- background noise
- language model
- speech processing
- speech synthesis
- noise reduction
- pattern recognition
- digit recognition
- speech recognition technology
- speech recognizer
- speech understanding
- noisy speech
- speaker identification
- speech enhancement
- speech recognition errors
- speech retrieval
- isolated word
- computer vision
- noise level
- signal to noise ratio
- neural network
- keyword spotting
- speaker independent
- speaker dependent
- noisy images
- signal processing
- feature extraction
- audio visual speech recognition