Component separation in astronomical images using independent factor analysis

Astronomical microwave images carry information about radiation from various sources, including cosmic microwave background radiation, galactic dust, synchrotron, etc. Moreover, the observations are corrupted with sensor noise, which is normally space varying. All of these components carry important information about the Universe and need to be recovered separately. In this paper, we study the problem of component separation in astronomical images using a recently introduced technique called independent factor analysis (IFA). We briefly analyse the source distributions and suggest a Gaussian mixture model. We then introduce IFA and discuss how the developed source and noise models are incorporated in the IFA algorithm. We present simulation results obtained by IFA on realistic data, which simulate the ones expected from the Planck Surveyor Satellite mission, to be launched by the European Space Agency.