Far-Field Speech Enhancement Using Heteroscedastic Autoencoder for Improved Speech Recognition.
Shashi KumarShakti P. RathPublished in: INTERSPEECH (2019)
Keyphrases
- speech recognition
- speech enhancement
- speech signal
- noisy environments
- noisy speech
- speech synthesis
- language model
- automatic speech recognition
- hidden markov models
- spectral subtraction
- vocal tract
- pattern recognition
- background noise
- speaker identification
- speech recognition systems
- linear prediction
- noise reduction
- additive noise
- signal to noise ratio
- machine learning
- smoothing algorithm
- single channel
- acoustic features
- spectral analysis
- neural network