Improved sparse representation based super-resolution

In this paper, super-resolution image is obtained from a single low-resolution image using dictionary learning approach. The original image is blurred and downsampled to the low-resolution image, and has to find the value which is lost during downsampling and trained with patches. Each patches of low-resolution image use that value of their respective high-resolution image during training of dictionary. The Hilbert phase congruency which provides more features and good edges and applied to each patches. Then, LR and HR patches of the dictionary are used to generate the high-resolution image patch. In our approach, which results in good quality HR image and having better PSNR values than the other similar SR methods.