SERIL: Noise Adaptive Speech Enhancement using Regularization-based Incremental Learning.
Chi-Chang LeeYu-Chen LinHsuan-Tien LinHsin-Min WangYu TsaoPublished in: CoRR (2020)
Keyphrases
- incremental learning
- speech enhancement
- noisy environments
- noise reduction
- signal to noise ratio
- wiener filter
- single channel
- linear prediction
- noisy speech
- background noise
- speech signal
- incremental learning algorithm
- additive noise
- speech recognition
- supervised learning
- learning process
- vocal tract
- edge preserving
- image restoration
- semi supervised
- adaptive filtering
- spatial domain
- edge detection
- markov random field
- maximum likelihood
- probabilistic model
- pairwise
- prior information
- face recognition