Handling derivative filterbank features in bounded-marginalization-based missing data automatic speech recognition.
Marco KühnePublished in: INTERSPEECH (2015)
Keyphrases
- missing data
- automatic speech recognition
- filter bank
- missing values
- speech recognition
- low rank
- structure from motion
- speech signal
- incomplete data
- subband
- spectral analysis
- hidden markov models
- noisy data
- feature set
- image processing
- multiple imputation
- signal processing
- multiresolution
- feature extraction
- support vector machine
- shape analysis
- feature vectors
- pattern recognition
- multiscale
- conversational speech
- neural network
- data imputation