Autoregressive Articulatory WaveNet Flow for Speaker-Independent Acoustic-to-Articulatory Inversion.
Narjes BozorgMichael T. JohnsonMohammad SoleymanpourPublished in: SpeD (2021)
Keyphrases
- speech recognition
- speaker independent
- autoregressive
- speaker dependent
- speech recognition systems
- acoustic models
- non stationary
- hidden markov models
- speech signal
- moving average
- speech synthesis
- speech recognizer
- automatic speech recognition
- gaussian markov random field
- language model
- random fields
- speaker identification
- pattern recognition
- noisy environments
- sar images
- spectral analysis
- acoustic features
- image reconstruction
- vocal tract
- edge detection
- co occurrence
- similarity measure
- feature extraction
- feature selection
- computer vision
- neural network