Single Image Super Resolution Using Sparse Representation with Image Fusion Principle

Abstract

A suggested  single image super resolution algorithm is proposed based on image fusion principle. Magnetic resonance and computed tomography images are interpolated using two algorithms that use sparse-representation modeling with dictionary learning. The MR and CT images are fused either by discrete wavelet or curvelet transforms, then the fused result are interpolated by the same algorithms. Simulation results show that the fused  super resolution image provides higher PSNR values than the original CT and MR images by using these interpolation algorithms. Also experimentally we deduce that using the curvelet fusion technique provides better results than using the wavelet, and scaling-up by one of the sparse representation algorithm gives more better results than bicubic and the other sparse representation algorithms in almost all images.