Filterbank-based feature extraction for speech recognition and its application to voice mail transcription.
Jun HuangMukund PadmanabhanPublished in: INTERSPEECH (2000)
Keyphrases
- speech recognition
- filter bank
- feature extraction
- speech recognition technology
- speech recognition systems
- mel frequency cepstral coefficients
- handwriting recognition
- frequency domain
- subband
- speech recognition errors
- image coding
- speaker identification
- speech synthesis
- speech signal
- pattern recognition
- cepstral coefficients
- voice activity detection
- hidden markov models
- feature vectors
- spectral analysis
- multiresolution
- wavelet transform
- multiscale
- automatic speech recognition
- language model
- speaker recognition
- signal processing
- speech processing
- speech recognizer
- feature selection
- image compression
- noisy environments
- face recognition
- image processing
- computationally efficient
- speaker independent
- wavelet bases
- extracted features
- texture features
- extracting features
- text to speech
- similarity measure
- speaker diarization
- students with learning disabilities
- speaker dependent
- machine learning
- computational complexity
- acoustic features
- speaker verification
- feature space
- gaussian mixture model
- wavelet coefficients
- principal component analysis
- feature set
- bit rate