A Training and Inference Strategy Using Noisy and Enhanced Speech as Target for Speech Enhancement without Clean Speech.
Li-Wei ChenYao-Fei ChengHung-Shin LeeYu TsaoHsin-Min WangPublished in: INTERSPEECH (2023)
Keyphrases
- speech enhancement
- noisy environments
- noisy speech
- speech signal
- speech recognition
- noise reduction
- signal to noise ratio
- vocal tract
- single channel
- background noise
- automatic speech recognition
- hidden markov models
- linear prediction
- speaker identification
- smoothing algorithm
- additive noise
- speech synthesis
- bayesian networks
- text to speech
- speaker recognition
- broadcast news
- acoustic features
- spectral analysis
- noisy images
- language model
- edge detection
- information retrieval