Industrial Noisy Speech Enhancement Using Joint Time-Frequency Loss Function Based on U-Net.
Rongxin QinZhigang LianPublished in: 6GN (2) (2023)
Keyphrases
- loss function
- speech enhancement
- noisy environments
- noise reduction
- pairwise
- speech signal
- speech recognition
- background noise
- support vector
- single channel
- signal to noise ratio
- automatic speech recognition
- frequency domain
- linear prediction
- signal processing
- least squares
- wiener filter
- wavelet transform
- machine learning
- smoothing algorithm
- image processing