Zoom-based super-resolution reconstruction approach using prior total variation

We present a robust and efficient approach for zoom-based super-resolution (SR) reconstruction problems. We employ the total variation (TV) of the desired image priori in the maximum a-posteriori estimation. An efficient algorithm based on iterative methods and preconditioning techniques is employed to solve the resulting variational problem. To suit the proposed algorithm for realistic imaging situations, a registration method is presented to simultaneously solve the zooming factors, image center shifts, and photometric parameters. Experimental results show that the proposed TV-based algorithm performs quite well in terms of both quantitative measurements and visual evaluation. We also demonstrate that the proposed algorithm is robust for SR image inpainting, where some pixels are missed in the SR reconstruction model.