An unsupervised framework jointly trains separation and direction-of-arrival networks by maximizing the evidence lower bound of a complex Gaussian mixture model, then uses the trained network to initialize multichannel EM separation.
This training is based on the LDA model [9, 24], which has TF masks and DoAs of sources as latent variables
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SD 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Deep Bayesian Unsupervised Source Separation Based on a Complex Gaussian Mixture Model
An unsupervised framework jointly trains separation and direction-of-arrival networks by maximizing the evidence lower bound of a complex Gaussian mixture model, then uses the trained network to initialize multichannel EM separation.