Elsevier, Pattern Recognition Letters, 4(27), p. 287-293, 2006
DOI: 10.1016/j.patrec.2005.08.010
Full text: Download
We present a maximum a posteriori solution to the problem of obtaining a high-resolution image from a set of degraded low-resolution images of the same scene. The proposed algorithm has the advantage that no prior knowledge of blurring functions is required and it can handle unknown misregistrations between the input images. An efficient implementation scheme of alternating minimizations is presented together with experiments that demonstrate the performance of the algorithm.