Weighted Likelihood Ratio (WLR) Hidden Markov Model for Noisy Speech Recognition.
Chao HuangYingchun HuangFrank K. SoongJianlai ZhouPublished in: ICASSP (1) (2006)
Keyphrases
- speech recognition
- hidden markov models
- likelihood ratio
- noisy environments
- hypothesis testing
- speech synthesis
- automatic speech recognition
- speech signal
- noisy speech
- speech processing
- speaker independent
- conditional random fields
- speech recognizer
- keyword spotting
- handwriting recognition
- speech recognition systems
- recognition engine
- computer vision
- digit recognition
- speaker identification
- feature extraction
- machine learning
- isolated word recognition