Entropy as a Stopping Metric for Deconvolving VLBI Images of AGN Jets

We present the results of our study of entropy as a potential stopping metric for the CLEAN algorithm that is used to deconvolve Very Long Baseline Interferometry (VLBI) images of radio jets from active galactic nuclei (AGN). Despite CLEAN's broad success in producing deconvolutions of VLBI images, determining the correct stopping point for the iterative algorithm to produce the most accurate images is still a challenge. Our goal was to find a stopping criteria based on the features of the residual image without prior knowledge of the jet structure or expected noise levels. We found that calculating entropy as a measure of randomness in the residual image has the potential to provide a reliable and automatic stopping criteria near its maximum value that produces deconvolved images of similar quality to those we were able to produce with a fixed stopping criteria set at some factor times the expected noise level