Generalized MAP estimation via parameter scheduling and maximizer of the posterior marginal estimate for image reconstruction using multiple halftone images

We studied performance of a generalized MAP estimation which was regarded as the maximizer of the posterior marginal (MPM) estimate for image reconstruction via multiple halftone images for a set of grayscale images generated by an assumed true prior and a 256-grayscale standard image. By making use of numerical simulations for those images, we clarified that performance of the generalized MAP estimation is improved with the increase in the number of the halftone images for image reconstruction, if we tune parameter scheduling appropriately. Also, we found that the present method reconstructs original images more accurately than the conventional MAP estimation and the MPM estimate.