Improvement in image resolution based on dispersive representation of data

Abstract

A method for reconstructing the resolution of images, based on selection and optimization of significant local features and sparse representation of processed-image blocks (using optimized low- and high-resolution dictionaries), has been substantiated for the first time. This method, making it possible to improve significantly the resolution of images of various nature, is interpreted physically. A block diagram of the processing system corresponding to the new approach to image reconstruction has been developed. A simulation of the new method for reconstructing images of different physical natures and known algorithms showed an advantage of the new scheme for reconstructing resolution in terms of universally recognized criteria (peak signal-to-noise ratio, mean absolute error, and structural similarity index measure) and in visual comparison of the processed images.