Toward image super-resolution based on local regression and nonlocal means

• Published in: Multimedia tools and applications Vol. 81; no. 16; pp. 23473 - 23492
• Main Authors: Bastanfard, Azam, Amirkhani, Dariush, Mohammadi, Mohammad
• Format: Journal Article
• Language: English
• Published: New York Springer US 01.07.2022; Springer Nature B.V
• Subjects: Closed circuit television; Computer Communication Networks; Computer engineering; Computer Science; Data Structures and Information Theory; Dictionaries; Genetic algorithms; Image reconstruction; Image resolution; Image retrieval; Methods; Multimedia; Multimedia Information Systems; Pattern recognition; Satellite imagery; Sensors; Special Purpose and Application-Based Systems; Single image super-resolution (SISR); Local regression; Nonlocal means; Super-resolution (SR)
• ISSN: 1380-7501; 1573-7721
• DOI: 10.1007/s11042-022-12584-x

Super-resolution algorithms have many weaknesses because they have low resolution and quality. These weaknesses are due to limitations and high cost of hardware implementation. Therefore, software methods are used due to its high accuracy and low cost. Super-resolution is used in a variety of applications, including satellite imagery, medicine, CCTV cameras, etc. Super-resolution is divided into two main categories: super-resolution using an input image and super-resolution using multiple input images. Considering the great use of the subject, this paper provides a method for increasing the resolution using an input image. In the proposed method, we improved the image super-resolution based on local regression and nonlocal means. The proposed method consists of two stages: a learning phase and a reconstruction phase. After the initial implementation, we compared the results with a quick and accurate method and found the difference. Then, using the Genetic algorithm, these differences were applied to the initial image. Results showed that our proposed method performed 2.4% and 2.2% better than the RAISR and LANR-NLM methods, Respectively.