A super-resolution method based on hybrid of generalized PMAP and POCS

The Poisson MAP (PMAP) and projection of convex sets (POCS) are two kinds of important image super-resolution (SR) reconstruction method. In this paper, we generalize the classic Poisson MAP (PMAP) method according to more general imaging model which contains down-sampling and sub-pixel shift operator. The Generalized PMAP (GPMAP) expands the applied scope of PMAP and removes ring artifacts effectively. A high resolution image reconstruction method based on hybrid of GPMAP and POCS is then proposed. The qualitative and quantitative analysis of experiment results demonstrates that the proposed hybrid algorithm is superior to the single GPMAP or POCS algorithm.