Studying the Effects of Masking and Deconvolution Algorithms in Imaging

Usually when observers get their data from a radio interferometer, they synthesize an image using a deconvolution algorithm they know from experience. Instead, we would like to formally quantify which algorithms might be preferable in certain circumstances, and specifically when masking might be necessary. With four datasets, we have many trials in which we change specific parameters such as: deconvolution algorithm, iterations, number of visibilities (interferometer responses), number of sources, masking, and threshold. Masking helps in situations where the algorithm doesn't have enough information to find a solution image. In this case, if a mask isn't specified, divergence is likely. However, ultimately on a systematic level, it matters more with which deconvolution algorithm one uses. While there are many different types of deconvolution algorithms, we've focused on Hogbom, Multi-scale, and Maximum Entropy Method (MEM).