Align-Denoise: Single-Pass Non-Autoregressive Speech Recognition.
Nanxin ChenPiotr ZelaskoLaureano Moro-VelázquezJesús VillalbaNajim DehakPublished in: Interspeech (2021)
Keyphrases
- speech recognition
- single pass
- autoregressive
- denoising
- non stationary
- hidden markov models
- language model
- random fields
- speech recognizer
- speech synthesis
- automatic speech recognition
- noisy environments
- spectral analysis
- image denoising
- pattern recognition
- speech signal
- speech recognition technology
- sar images
- speech recognition systems
- speaker identification
- image fusion
- natural images
- machine learning
- motion estimation
- linear prediction
- information extraction
- speaker independent