Maximum Likelihood Imaging for Sensor ArraysPP

We propose the Sieved Maximum Likelihood (SiML) method for imaging purposes in array signal processing. It uses a general functional data model, allowing for an unrestricted number of arbitrarily-shaped sources. As the likelihood problem elicits many solutions, we restrict the search space dimension using an information criterion. Asymptotic confidence intervals are derived. We show that SiML is computationally very efficient, and results in much better accuracy than the traditional spectral-based methods.

This is poster number 36 in Poster Session

Matthieu Simeoni (EPFL / IBM Research)
Paul Hurley (IBM Research, Zurich)
array signal processing, image reconstruction, inverse problems, random fields, sampling operator, sieved maximum likelihood, statistical inverse estimation methods, stochastic processes