Super resolution based on simultaneous registration and reconstruction

The progress of super-resolution usually involves computing the parameters (e.g. the registration and blur parameters) of a generating imaging model, and obtaining an estimated SR image by minimizing a cost function. In this paper, a high-resolution image, the registration parameters and PSF width will be optimized simultaneously using scaled conjugate gradient. A pre-registration method is introduced to ensure convergence and save time. The experiment results show that this method is valid.