Multi-fidelity and learning-regularization for single image super resolution (Q2148464)

The new approach to super-resolution ill-posed problem is proposed using regularization. A~multi-fidelity term is employed to measure the similarities between the observed data and the simulated data. The restoration procedure is decomposed into regularization-based super-resolution problem's solver and a denoising problem's solver. The effectiveness of the proposed method is validated using both synthetic and real-world low resolution test images.