Bitstream-based feature extraction for wireless speech recognition.
Hong Kook KimRichard V. CoxPublished in: ICASSP (2000)
Keyphrases
- speech recognition
- bitstream
- feature extraction
- coding scheme
- bit rate
- pattern recognition
- cepstral coefficients
- speaker identification
- video quality
- compression algorithm
- video transmission
- automatic speech recognition
- wireless communication
- variable length
- hidden markov models
- language model
- speech synthesis
- wireless channels
- speech signal
- speech recognizer
- scalable video coding
- compressed domain
- image processing
- error concealment
- speech recognition technology
- wavelet coefficients
- feature vectors
- image classification
- video coding
- frequency domain
- principal component analysis
- mel frequency cepstral coefficients
- noisy environments
- speech recognition systems
- wavelet transform
- speaker independent
- feature selection
- rate distortion
- mobile devices
- extracting features
- feature space
- face recognition
- computational complexity
- machine learning
- speaker recognition
- multiresolution
- image quality
- subband
- prediction error
- motion vectors