Improving Noise Robust Automatic Speech Recognition with Single-Channel Time-Domain Enhancement Network.
Keisuke KinoshitaTsubasa OchiaiMarc DelcroixTomohiro NakataniPublished in: ICASSP (2020)
Keyphrases
- noisy environments
- automatic speech recognition
- speech enhancement
- single channel
- speech recognition
- speech signal
- multi channel
- frequency domain
- noise reduction
- conversational speech
- broadcast news
- wiener filter
- additive noise
- independent component analysis
- non stationary
- hidden markov models
- image processing
- neural network
- signal to noise ratio
- multi modal
- linear prediction
- sound source
- multiscale
- machine learning