Ripple sparse self-attention for monaural speech enhancement.
Qiquan ZhangHongxu ZhuQi SongXinyuan QianZhaoheng NiHaizhou LiPublished in: CoRR (2023)
Keyphrases
- speech enhancement
- noisy environments
- noise reduction
- single channel
- signal to noise ratio
- speech signal
- high dimensional
- wiener filter
- smoothing algorithm
- multi channel
- linear prediction
- pattern recognition
- speech recognition
- hidden markov models
- image processing
- automatic speech recognition
- background noise
- image segmentation