EMGSE: Acoustic/EMG Fusion for Multimodal Speech Enhancement.
Kuan-Chen WangKai-Chun LiuHsin-Min WangYu TsaoPublished in: ICASSP (2022)
Keyphrases
- speech enhancement
- noisy speech
- noisy environments
- noise reduction
- signal to noise ratio
- speech signal
- speech recognition
- single channel
- hidden markov models
- acoustic features
- linear prediction
- vocal tract
- multi modal
- background noise
- audio visual
- smoothing algorithm
- wiener filter
- multiscale
- edge detection
- edge preserving
- signal processing
- image quality
- video sequences