Audio source separation using a redundant library of source spectral bases for non-negative tensor factorization.
Mahmoud FakhryPiergiorgio SvaizerMaurizio OmologoPublished in: ICASSP (2015)
Keyphrases
- tensor factorization
- source separation
- audio features
- blind source separation
- negative matrix factorization
- matrix factorization
- independent component analysis
- collaborative filtering
- singular value decomposition
- audio visual
- tag recommendation
- multimedia
- signal processing
- visual features
- feature set
- denoising
- low level
- principal component analysis
- music information retrieval
- visual information
- sparse representation
- multi modal
- co occurrence
- feature vectors