Sensitivity analysis of correspondence ghost imaging

Abstract

In this letter, sensitivity analysis is adopted to interpret why correspondence ghost imaging can restore better images with fewer reference patterns which have larger bucket value fluctuation: these selected patterns are more alike with regard to the target to be restored. The contribution of reference patterns to imaging can be quantified as sensitivity coefficients. In addition, the peak signal-to-noise ratio and mean square error are calculated to compare the restored image quality, and results confirmed that sensitivity coefficients are more robust than bucket value fluctuation in the presence of noise for selecting reference patterns.