A perceptually-weighted deep neural network for monaural speech enhancement in various background noise conditions.
Qingju LiuWenwu WangPhilip J. B. JacksonYan TangPublished in: EUSIPCO (2017)
Keyphrases
- background noise
- speech enhancement
- neural network
- noisy environments
- speech signal
- low contrast
- noisy speech
- noise reduction
- speech recognition
- signal to noise ratio
- non stationary
- automatic speech recognition
- adaptive filtering
- pattern recognition
- multiscale
- vocal tract
- human visual system
- multi modal
- super resolution
- linear prediction
- single channel
- bayesian networks
- computer vision