Bayesian blind component separation for Cosmic Microwave Background observations
arXiv:astro-ph/0109123 · doi:10.1063/1.1477043
Abstract
We present a technique for the blind separation of components in CMB data. The method uses a spectral EM algorithm which recovers simultaneously component templates, their emission law as a function of wavelength, and noise levels. We test the method on Planck HFI simulated observations featuring 3 astrophysical components.
15 pages, 5 figures, to appear in the Proceedings of the MAXENT 2001 international workshop