Efficient learning based Image Super Resolution via Sparse Representation

Abstract

This paper addresses the problem of generating a superresolution (SR) image from a single low-resolution input image. Research on image statistics suggests that image patches can be well represented as a sparse linear combination of elements from an appropriately chosen over-complete dictionary. Inspired by this observation, we seek a sparse representation for each patch of the low-resolution input, and then use the coefficients of this representation to generate the high-resolution output. By jointly training two dictionaries for the low- and highresolution image patches, we can enforce the similarity of sparse representations between the low resolution and high resolution image patch pair with respect to their own dictionaries. Therefore, the sparse representation of a low resolution image patch can be applied with the high resolution image patch dictionary to generate a high resolution image patch. The learned dictionary pair is a more compact representation of the patch pairs, compared to previous approaches, which simply sample a large amount of image patch pairs [1], reducing the computational cost substantially. The effectiveness of such sparsity prior is demonstrated for general image superresolution. In this case, our algorithm generates high-resolution images that are competitive or even superior in quality to images produced by other similar SR methods. In addition, the local sparse modeling of our approach is naturally robust to noise, and therefore the proposed algorithm can handle super-resolution with noisy inputs in a more unified framework.

Authors and Affiliations

R. Sudheer Babu, Dr. K. E. Sreenivasa Murthy

Keywords

Related Articles

A design for processors using lasers to process data

Laser employing microprocessors can significantly increase clock speed, as it would operate with lasers working in successive patterns of on and off, unlike current processors, which use electricity to transmit pulses of...

Analysis On High Rate GNSS Data Co-Seismic Response In Jiuzhaigou 7.0Ms Earthquake

Jiuzhaigou 7.0Ms Earthquake Occured At 13:19 8th August 2017 (UTC) In Sichuan Province, China. We Processed The High Rate (1Hz) Global Navigation Satellite System(GNSS) Data Of Stations In TwoHundred Kilometers Distance...

Study of the Induction of Secondary Growth of Alcoholic Yeast Saccharomyces Cerevisiae

The results of the study of the effect of MgSO4, CuSO4 and ultraviolet irradiation on the induction of secondary growth of the colonies of alcohol yeast Saccharomyces cerevisiae are presented. It was found that secondary...

The Effects of Heat on Electronic Components

Regardless of the cause of overheating, the general and primary effect of overheating on an electronic component is damage. In this paper, the effects of excessive levels heat on different types of electronic components...

Tourism and Economic Development – A Case Study In Tamil Nadu

Tourism industry offers much scope for earning foreign exchange and it stimulates the rate of growth of the economy. Through interaction of natural, human and cultural factors, it functions as a major means of recreation...

Download PDF file
  • EP ID EP392014
  • DOI 10.9790/9622-0708033949.
  • Views 86
  • Downloads 0

How To Cite

R. Sudheer Babu, Dr. K. E. Sreenivasa Murthy (2017). Efficient learning based Image Super Resolution via Sparse Representation. International Journal of engineering Research and Applications, 7(8), 39-49. https://europub.co.uk/articles/-A-392014