Monophonic sound source separation by non-negative sparse autoencoders.
Keiki ZenMasahiro SuzukiHaruhiko SatoSatoshi OyamaMasahito KuriharaPublished in: SMC (2014)
Keyphrases
- source separation
- mixing matrix
- denoising
- musical instrument
- blind source separation
- musical instruments
- independent component analysis
- audio features
- audio signal
- music information retrieval
- single channel
- temporal structure
- sparse representation
- image processing
- simple linear
- sound source
- music retrieval
- high dimensional
- multiresolution
- random projections
- negative matrix factorization
- wiener filter
- factor analysis
- audio visual
- low level
- image retrieval
- face recognition
- signal processing