Measured and simulated passive millimeter-wave images of a simple target pattern were processed with Two-Mu, a linear, noniterative restoration algorithm. Decisions about whether the RAM pattern objects were sufficiently visible against a metallic background to be recognized were made by visually inspecting the images. This allowed quantitative measures to be made of the resolution improvement afforded by Two-Mu, as a function of scene oversampling for various target ranges and radiometer apertures. Improvements by factors up to 1.25 and 1.45 were found for images and analog plots, respectively, of cuts through images, respectively.