Super-resolution processing using luminance dispersion in local area

Abstract

Based on super-resolution processing, reconstruction refers to the technique of transforming several low-resolution images into a high-resolution image as a general tendency. The fractal super-resolution method is the most effective technique for obtaining a high-resolution image from a single low-resolution image. This technology performs expansion by partitioning the original image into block groups in the local region, which are called range blocks, and block groups in the global region, which are called domain blocks. Further, it replaces a range block with the domain block that resembles it the most. In this study, blocks containing image edges are subjected to fractal super-resolution processing using luminance variance, and areas with gradual changes in density are expanded by interpolation.