L2-Boosting-based dictionary learning for super-resolution

Based on the assumption of sparse representation and the theory of compressed sensing, Yang et al. propose an excellent super-resolution algorithm. However, the process of training coupled dictionaries cannot be perfectly connected with the process of reconstructing super-resolution images in theory. Therefore, a novel dictionary-based super-resolution algorithm is proposed in this paper. Different from Yang´s algorithm, the low- and high-resolution dictionaries are separately trained by employing an L₂-Boosting algorithm. Extensive experiments validate that our algorithm can surpass Yang´s algorithm in both visual perception and statistical performance.