Detection Methods of Salient Regions in Super-Resolution Based on Sparse Representation

Achieving successful results with the sparse representation in super-resolution increases the interest in the field. The sparse representation model, which is an important method in super-resolution, consists of image patches, a correct dictionary and a sparse linear combination of the elements of this dictionary. At this point, the super-resolution successfully reflects the sparse pattern by obtaining high-resolution images with the sparse pattern from low-resolution image patches. The detection of image regions is critical here. In the proposed method, the successes of the results are compared by using Fuzzy C-Means Clustering and Hue-Saturation-Value (HSV) Based Segmentation methods for determination of these regions.