Bayesian Non-parametric Multi-source Modelling Based Determined Blind Source Separation

Abstract

This paper proposes a determined blind source separation method using Bayesian non-parametric modelling of sources. Conventionally source signals are separated from a given set of mixture signals by modelling them using non-negative matrix factorization (NMF). However in NMF, a latent variable signifying model complexity must be appropriately specified to avoid over-fitting or under-fitting. As real-world sources can be of varying and unknown complexities, we propose a Bayesian non-parametric framework which is invariant to such latent variables. We show that our proposed method adapts to different source complexities, while conventional methods require parameter tuning for optimal separation.

A Unified Bayesian Source Modelling for Determined Blind Source Separation

“…It is evident that the proposed method is able to outperform both ILRMA and U-ILRMA. Further, ILRMA seems to perform better with smaller K as opposed to U-ILRMA which prefers a higher K. We further compare our proposed method with the Bayesian generalization of ILRMA (Bay-ILRMA) [13] with K = 30 bases. The comparison of overall SIR, SAR and SDR are summarized in Table. 2.…”

“…In such techniques, it is common to introduce hidden variables to capture the structure of given observed data, and then inference them to estimate the posterior distribution. We recently proposed a Bayesian generalization of ILRMA [13].…”

“…This motivates us to reformulate these methods with more flexible frameworks. We recently proposed a Bayesian generalization for the first ILRMA approach and showed that it can adaptively model a diverse range of sources [13]. However, a unified multi-source model can capture the correlations among sources, which is not possible when each source is modelled separately.…”