Single-channel speech separation based on NLMSE quasi-KLT bases.
Haiyan GuoZhen YangWeiping ZhuPublished in: WCSP (2011)
Keyphrases
- single channel
- speech enhancement
- sound source
- multi channel
- speech signal
- audio visual
- speech recognition
- independent component analysis
- frequency domain
- prior information
- wiener filter
- noise reduction
- multi modal
- brain activity
- vocal tract
- noisy environments
- automatic speech recognition
- wavelet decomposition
- non stationary
- linear prediction
- computer vision
- signal to noise ratio
- image segmentation