SRTNET: Time Domain Speech Enhancement via Stochastic Refinement.
Zhibin QiuMengfan FuYinfeng YuLiLi YinFuchun SunHao HuangPublished in: ICASSP (2023)
Keyphrases
- speech enhancement
- noisy environments
- noise reduction
- single channel
- signal to noise ratio
- frequency domain
- speech signal
- linear prediction
- vocal tract
- wiener filter
- smoothing algorithm
- multi channel
- additive noise
- subband
- pattern recognition
- high frequency
- speech recognition
- edge detection
- multiresolution
- feature vectors