Data-driven temporal filters based on multi-eigenvectors for robust features in speech recognition.
Ni-Chun WangJeih-Weih HungLin-Shan LeePublished in: ICASSP (1) (2003)
Keyphrases
- speech recognition
- data driven
- speech recognition systems
- noisy environments
- hidden markov models
- speech synthesis
- language model
- automatic speech recognition
- cepstral coefficients
- pattern recognition
- speech recognizer
- speech understanding
- speech signal
- feature set
- feature space
- speech processing
- neural network
- speech recognition technology
- feature extraction
- edge detection
- speech recognition errors
- low level
- feature vectors
- speaker independent
- extracted features
- speech enhancement
- extracting features
- principal component analysis
- image processing