A combination of discriminative and maximum likelihood techniques for noise robust speech recognition.
Kari LaurilaMarcel VasilacheOlli ViikkiPublished in: ICASSP (1998)
Keyphrases
- speech recognition
- noisy environments
- maximum likelihood
- speaker adaptation
- speech signal
- additive noise
- hidden markov models
- speech processing
- language model
- speech synthesis
- speech recognition technology
- automatic speech recognition
- speech understanding
- pattern recognition
- speech recognizer
- noisy speech
- background noise
- discriminative training
- speech enhancement
- noise reduction
- speaker independent
- speaker identification
- digit recognition
- speech recognition systems
- word recognition
- model selection
- speech recognition errors
- speech retrieval
- feature extraction
- information retrieval
- expectation maximization