Image Super-Resolution Based on Sparse Coding with Multi-Class Dictionaries

keywords: Image patch classification, multi-class dictionaries, phase congruency, sparse coding, super-resolution
Sparse coding-based single image super-resolution has attracted much interest. In this paper, a super-resolution reconstruction algorithm based on sparse coding with multi-class dictionaries is put forward. We propose a novel method for image patch classification, using the phase congruency information. A sub-dictionary is learned from patches in each category. For a given image patch, the sub-dictionary that belongs to the same category is selected adaptively. Since the given patch has similar pattern with the selected sub-dictionary, it can be better represented. Finally, iterative back-projection is used to enforce global reconstruction constraint. Experiments demonstrate that our approach can produce comparable or even better super-resolution reconstruction results with some existing algorithms, in both subjective visual quality and numerical measures.
reference: Vol. 38, 2019, No. 6, pp. 1301–1319