Inverse Dithering through IMAP Estimation

摘要

In this paper we extend the use of the IMAP (Indirect Maximum A Priori) estimation to the inverse dithering problem. To find a continuous grey-levels image from a dithered image we have to estimate an appropriate blur mask. By the IMAP estimation we find the blur mask that permits to obtain the image that better fits the a priori knowledge on it. To reduce the computational cost we use an E-GNC (Extended Graduated Non-Convex) algorithm to perform the image restoration during the estimation.