End-To-End Speech Recognition with Joint Dereverberation of Sub-Band Autoregressive Envelopes.
Rohit KumarAnurenjan PurushothamanAnirudh SreeramSriram GanapathyPublished in: ICASSP (2022)
Keyphrases
- end to end
- speech recognition
- autoregressive
- subband
- filter bank
- wavelet transform
- multiresolution
- discrete wavelet transform
- high frequency
- wavelet decomposition
- non stationary
- image compression
- hidden markov models
- bit rate
- wavelet coefficients
- wavelet domain
- frequency domain
- random fields
- low frequency
- feature vectors
- language model
- pattern recognition
- sar images
- speech signal
- frequency band
- multiscale
- wavelet packet
- speaker identification
- image analysis
- linear prediction
- visual quality
- image coding
- higher order