Unsupervised Single-Channel Separation of Nonstationary Signals Using Gammatone Filterbank and Itakura-Saito Nonnegative Matrix Two-Dimensional Factorizations.
Bin GaoWai Lok WooSatnam Singh DlayPublished in: IEEE Trans. Circuits Syst. I Regul. Pap. (2013)
Keyphrases
- single channel
- non stationary
- nonnegative matrix
- nonnegative matrix factorization
- filter bank
- bregman divergences
- frequency domain
- sound source
- multi channel
- signal processing
- subband
- relative entropy
- speech signal
- spectral analysis
- wavelet decomposition
- multiresolution
- independent component analysis
- negative matrix factorization
- matrix factorization
- data representation
- principal component analysis
- multiscale
- least squares
- spatial domain
- brain activity
- autoregressive
- prior information
- semi supervised
- random fields
- wiener filter
- spectral clustering
- feature extraction
- computationally efficient
- image quality
- image compression
- blind source separation
- denoising
- wavelet transform
- document clustering
- graphical models
- markov random field
- clustering method
- wavelet coefficients
- wavelet domain