Abstract

In this letter, we generalize the convolutional NMF by taking the$\beta$-divergence as the contrast function and present the correctmultiplicative updates for its factors in closed form. The new updates unifythe $\beta$-NMF and the convolutional NMF. We state why almost all of theexisting updates are inexact and approximative w.r.t. the convolutional datamodel. We show that our updates are stable and that their convergenceperformance is consistent across the most common values of $\beta$.

 

Quick Read (beta)