Data-driven temporal filters based on maximum mutual information for robust features in speech recognition.
Yung-Sheng HuangJeih-Weih HungPublished in: ISCSLP (2004)
Keyphrases
- speech recognition
- data driven
- mutual information
- speech recognition systems
- noisy environments
- hidden markov models
- speech processing
- language model
- cepstral coefficients
- image registration
- pattern recognition
- speech synthesis
- feature set
- speech signal
- automatic speech recognition
- probabilistic model
- low level
- gabor filters
- feature space
- speaker identification
- speaker dependent
- computer vision
- speech recognizer
- signal processing
- similarity measure