Single-channel speech separation using sparse non-negative matrix factorization.
Mikkel N. SchmidtRasmus Kongsgaard OlssonPublished in: INTERSPEECH (2006)
Keyphrases
- matrix factorization
- sparse non negative
- single channel
- sound source
- speech signal
- multi channel
- sparse coding
- audio visual
- collaborative filtering
- recommender systems
- speech recognition
- missing data
- independent component analysis
- factor analysis
- nonnegative matrix factorization
- negative matrix factorization
- frequency domain
- wiener filter
- wavelet decomposition
- signal processing
- prior information
- brain activity
- feature extraction
- high dimensional
- hidden markov models