GIRA

Another deficiency of plucy was that it made use of an entropy constraint on the background. This meant that the user had to choose values for two free parameters: the ``smoothing kernel'' for the background and the ``coupling constant" which controlled how strongly the entropy term contributed to the objective function which was being maximised. The complexity of this two-term method also could lead to convergence problems under some circumstances. Leon Lucy therefore developed an alternative approach which achieved similar ends within a pure maximum likelihood framework. In this case the background channel is represented not as a set of independent pixels but as the sum of a set of Gaussians, one for each pixel. The width of these Gaussians is a free parameter, the only one, which constrains the maximum fine structure in the background in an intuitively simple manner. This method also has the potential to optimise the positions of the point-sources as well as their flux. The implementation of this method, called gira, is well advanced and is described in a companion paper by Pirzkal et al. in this newsletter.