Regularized Multiframe Super-resolution Model

The function implements the regularized multiframe super-resolution model

This function implements the multiframe super-resolution model based on the diffusion-driven regularizing functional (regularizeSR.m). The inputs to the function are: s (input image sequences), delta_est (translational motion values in the horizontal and vertical directions), phi_est (rotational motion), and factor (magnification). The registration method that estimates motions (rotations and translations) is that of Keren (Please see the work of Iran and colleagues, "Improving resolution by image registration"). The corresponding paper, entitled "A Noise-Suppressing and Edge-Preserving Multi-Frame Super-Resolution Image Reconstruction Method", has been submitted to the Journal of Signal Processing: Image Communication (Manuscript ID: image3133) and is in its second revision. Note that the super-resolution GUI, available at http://lcav.epfl.ch/software/superresolution, can easily be modified to include this function.