Fundamental limitation of frequency domain blind source separation for convolutive mixture of speech.
Shoko ArakiShoji MakinoTsuyoki NishikawaHiroshi SaruwatariPublished in: ICASSP (2001)
Keyphrases
- frequency domain
- blind source separation
- convolutive mixtures
- blind separation
- speech signal
- source separation
- fourier transform
- spatial domain
- speech recognition
- source signals
- denoising
- single channel
- independent component analysis
- independent components
- non stationary
- subband
- feature extraction
- filter bank
- low pass
- hidden markov models
- spectral analysis
- signal processing
- dimensionality reduction
- negative matrix factorization
- factor analysis
- eeg data
- wavelet transform
- image features
- audio visual
- object recognition
- computational complexity