Table of Contents Author Guidelines Submit a Manuscript
Mathematical Problems in Engineering
Volume 2015, Article ID 312423, 10 pages
http://dx.doi.org/10.1155/2015/312423
Research Article

Single Image Superresolution Using Maximizing Self-Similarity Prior

1National Engineering Research Center of Digital Life, Sun Yat-sen University, Guangzhou 510006, China
2Science and Technology Department, Zhongkai University of Agriculture and Engineering, Guangzhou 510225, China
3Beijing Key Laboratory of Multimedia and Intelligent Software Technology, College of Metropolitan Transportation, Beijing University of Technology, Beijing 100124, China
4School of Mathematics and Information Science, Nanchang Hangkong University, Nanchang 330063, China

Received 5 September 2015; Revised 10 November 2015; Accepted 12 November 2015

Academic Editor: Daniel Zaldivar

Copyright © 2015 Jianhong Li et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Single image superresolution (SISR) requires only one low-resolution (LR) image as its input which thus strongly motivates researchers to improve the technology. The property that small image patches tend to recur themselves across different scales is very important and widely used in image processing and computer vision community. In this paper, we develop a new approach for solving the problem of SISR by generalizing this property. The main idea of our approach takes advantage of a generic prior that assumes that a randomly selected patch in the underlying high-resolution (HR) image should visually resemble as much as possible with some patch extracted from the input low-resolution (LR) image. Observing the proposed prior, our approach deploys a cost function and applies an iterative scheme to estimate the optimal HR image. For solving the cost function, we introduce Gaussian mixture model (GMM) to train on a sampled data set for approximating the joint probability density function (PDF) of input image with different scales. Through extensive comparative experiments, this paper demonstrates that the visual fidelity of our proposed method is often superior to those generated by other state-of-the-art algorithms as determined through both perceptual judgment and quantitative measures.