A heuristic Bayesian design criterion for imaging resolution enhancement

In this theoretical study we propose an efficient, new approach to optimal experimental design (OED) for imaging resolution enhancement. We start by showing that designing experiments by seeking to minimize the forecast model uncertainty, as measured by the determinant of the posterior model covariance matrix, entails the maximization of the trace of the model resolution matrix. We, then, discuss the relevance of model resolution to imaging resolution and argue that the D-optimality criterion which minimizes the forecast model uncertainty also maximizes imaging resolution. The results are generic and may find applications in diverse fields such as medical imaging, microbiology, and geophysical exploration.