Single image super-resolution via internal gradient similarity

Abstract

Image super-resolution aims to reconstruct a high-resolution image from one or multiple low-resolution images which is an essential operation in a variety of applications. Due to the inherent ambiguity for superresolution, it is a challenging task to reconstruct clear, artifacts-free edges while still preserving rich and natural textures. In this paper, we propose a novel, straightforward, and effective single image superresolution method based on internal across-scale gradient similarity. The low-resolution gradients are first upsampled and then fed into an optimization framework to construct the final high-resolution output. The proposed approach is able to synthesize natural high-frequency texture details and maintain clean edges even under large scaling factors. Experimental results demonstrate that out method outperforms existing single image super-resolution techniques. We further evaluate the super-resolution performance when both internal statistics and external statistics are adopted. It is demonstrated that generally, internal statistics are sufficient for single image super-resolution.

Single image super-resolution using self-optimizing mask via fractional-order gradient interpolation and reconstruction

“…We compare our approach with the recent state-of-the-art algorithms which are explained in the following articles [23,34,35,24] in terms of commonly used super-resolution test examples which can be seen in the Berkeley dataset BSDS500. For methods based on gradient information we select the algorithms presented in articles [24,23] for comparison. For methods based on example-based information we chose algorithms presented in the articles [15,13,34]for comparison.…”

“…This method is discussed in the following articles, [16,17,18,19,20,21]. Given the similarity of the structure, gradient profile prior has got quite a bit of attention recently as seen in the following articles, [22,23,24,25,26]. Article [23] proposes an internal gradient similarity method which produces high-resolution image gradient samples which are used for further image reconstruction.…”

“…Given the similarity of the structure, gradient profile prior has got quite a bit of attention recently as seen in the following articles, [22,23,24,25,26]. Article [23] proposes an internal gradient similarity method which produces high-resolution image gradient samples which are used for further image reconstruction. Article [24] proposes using natural image gradients profile prior for image SR and the image gradient in this technique is modeled by using a parametric profile model.…”