Simple, Accurate, and Robust Nonparametric Blind Super-Resolution

This paper proposes a simple, accurate, and robust approach to single image nonparametric blind Super-Resolution (SR). This task is formulated as a functional to be minimized with respect to both an intermediate super-resolved image and a nonparametric blur-kernel. The proposed approach includes a convolution consistency constraint which uses a non-blind learning-based SR result to better guide the estimation process. Another key component is the unnatural bi-l0-l2-norm regularization imposed on the super-resolved, sharp image and the blur-kernel, which is shown to be quite beneficial for estimating the blur-kernel accurately. The numerical optimization is implemented by coupling the splitting augmented Lagrangian and the conjugate gradient (CG). Using the pre-estimated blur-kernel, we finally reconstruct the SR image by a very simple non-blind SR method that uses a natural image prior. The proposed approach is demonstrated to achieve better performance than the recent method by Michaeli and Irani [2] in both terms of the kernel estimation accuracy and image SR quality.