A Time-Frequency Attention Module for Neural Speech Enhancement.
Qiquan ZhangXinyuan QianZhaoheng NiAaron NicolsonEliathamby AmbikairajahHaizhou LiPublished in: IEEE ACM Trans. Audio Speech Lang. Process. (2023)
Keyphrases
- speech enhancement
- noisy environments
- neural network
- noise reduction
- signal to noise ratio
- speech signal
- linear prediction
- single channel
- signal processing
- frequency domain
- smoothing algorithm
- wavelet transform
- visual attention
- multiresolution
- focus of attention
- wiener filter
- motion estimation
- video sequences
- multiscale
- background noise
- image processing