Blind separation of noisy Gaussian stationary sources. Application to cosmic microwave background imaging

TL;DR

This paper introduces a likelihood-based source separation method tailored for noisy Gaussian stationary sources, specifically applied to cosmic microwave background imaging, leveraging spectral domain approximations and the EM algorithm.

Contribution

A novel spectral domain likelihood maximization approach for separating noisy Gaussian stationary sources, optimized for CMB anisotropy imaging.

Findings

Effective separation of Gaussian stationary sources in noisy conditions

Simplified likelihood form using spectral domain approximations

Successful application to cosmic microwave background imaging

Abstract

We present a new source separation method which maximizes the likelihood of a model of noisy mixtures of stationary, possibly Gaussian, independent components. The method has been devised to address the problem of imaging CMB anisotropies. It works in the spectral domain where, thanks to two simple approximations, the likelihood assumes a simple form which is easy to handle (low dimensional sufficient statistics) and to maximize (via the EM algorithm).