Abstract

Single image super-resolution (SR) aims at generating a plausible and visually pleasing high-resolution (HR) image from a low-resolution (LR) input. In this paper, we propose an effective single image SR algorithm by using collaborative representation and exploiting non-local self-similarity of natural images. In particular, the collaborative-representation-based method is applied to build the so-called self-projection matrices from a training set of HR images. Then the learned self-projection matrices are used to establish the collaborative-representation-based regularization (CRR), which is responsible for introducing the external HR information. Furthermore, to guarantee a reliable estimation of the HR image, the non-local low-rank regularization (NLR) which exploits internal prior of images is also taken into consideration. Since the CRR term and NLR term are complementary, they are assembled together to form a new reconstruction-based framework for SR recovery. Finally, to implement the proposed framework, an iterative algorithm is designed to gradually improve the quality of the SR results. Extensive experimental results indicate that the proposed approach is capable of delivering higher quality of SR results than several state-of-the-art SR methods.

Original languageEnglish (US)
Pages (from-to)49-61
Number of pages13
JournalSignal Processing
Volume149
DOIs
StatePublished - Aug 1 2018

Fingerprint

Image resolution
Optical resolving power
Recovery

Keywords

  • Collaborative representation
  • Non-local self-similarity
  • Regularization
  • Super-resolution

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Software
  • Signal Processing
  • Computer Vision and Pattern Recognition
  • Electrical and Electronic Engineering

Cite this

Single image super-resolution using collaborative representation and non-local self-similarity. / Chang, Kan; Ding, Pak Lun Kevin; Li, Baoxin.

In: Signal Processing, Vol. 149, 01.08.2018, p. 49-61.

Research output: Contribution to journalArticle

@article{00d6072799bd4307b77e921d56870628,
title = "Single image super-resolution using collaborative representation and non-local self-similarity",
abstract = "Single image super-resolution (SR) aims at generating a plausible and visually pleasing high-resolution (HR) image from a low-resolution (LR) input. In this paper, we propose an effective single image SR algorithm by using collaborative representation and exploiting non-local self-similarity of natural images. In particular, the collaborative-representation-based method is applied to build the so-called self-projection matrices from a training set of HR images. Then the learned self-projection matrices are used to establish the collaborative-representation-based regularization (CRR), which is responsible for introducing the external HR information. Furthermore, to guarantee a reliable estimation of the HR image, the non-local low-rank regularization (NLR) which exploits internal prior of images is also taken into consideration. Since the CRR term and NLR term are complementary, they are assembled together to form a new reconstruction-based framework for SR recovery. Finally, to implement the proposed framework, an iterative algorithm is designed to gradually improve the quality of the SR results. Extensive experimental results indicate that the proposed approach is capable of delivering higher quality of SR results than several state-of-the-art SR methods.",
keywords = "Collaborative representation, Non-local self-similarity, Regularization, Super-resolution",
author = "Kan Chang and Ding, {Pak Lun Kevin} and Baoxin Li",
year = "2018",
month = "8",
day = "1",
doi = "10.1016/j.sigpro.2018.02.031",
language = "English (US)",
volume = "149",
pages = "49--61",
journal = "Signal Processing",
issn = "0165-1684",
publisher = "Elsevier",

}

TY - JOUR

T1 - Single image super-resolution using collaborative representation and non-local self-similarity

AU - Chang, Kan

AU - Ding, Pak Lun Kevin

AU - Li, Baoxin

PY - 2018/8/1

Y1 - 2018/8/1

N2 - Single image super-resolution (SR) aims at generating a plausible and visually pleasing high-resolution (HR) image from a low-resolution (LR) input. In this paper, we propose an effective single image SR algorithm by using collaborative representation and exploiting non-local self-similarity of natural images. In particular, the collaborative-representation-based method is applied to build the so-called self-projection matrices from a training set of HR images. Then the learned self-projection matrices are used to establish the collaborative-representation-based regularization (CRR), which is responsible for introducing the external HR information. Furthermore, to guarantee a reliable estimation of the HR image, the non-local low-rank regularization (NLR) which exploits internal prior of images is also taken into consideration. Since the CRR term and NLR term are complementary, they are assembled together to form a new reconstruction-based framework for SR recovery. Finally, to implement the proposed framework, an iterative algorithm is designed to gradually improve the quality of the SR results. Extensive experimental results indicate that the proposed approach is capable of delivering higher quality of SR results than several state-of-the-art SR methods.

AB - Single image super-resolution (SR) aims at generating a plausible and visually pleasing high-resolution (HR) image from a low-resolution (LR) input. In this paper, we propose an effective single image SR algorithm by using collaborative representation and exploiting non-local self-similarity of natural images. In particular, the collaborative-representation-based method is applied to build the so-called self-projection matrices from a training set of HR images. Then the learned self-projection matrices are used to establish the collaborative-representation-based regularization (CRR), which is responsible for introducing the external HR information. Furthermore, to guarantee a reliable estimation of the HR image, the non-local low-rank regularization (NLR) which exploits internal prior of images is also taken into consideration. Since the CRR term and NLR term are complementary, they are assembled together to form a new reconstruction-based framework for SR recovery. Finally, to implement the proposed framework, an iterative algorithm is designed to gradually improve the quality of the SR results. Extensive experimental results indicate that the proposed approach is capable of delivering higher quality of SR results than several state-of-the-art SR methods.

KW - Collaborative representation

KW - Non-local self-similarity

KW - Regularization

KW - Super-resolution

UR - http://www.scopus.com/inward/record.url?scp=85043605701&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85043605701&partnerID=8YFLogxK

U2 - 10.1016/j.sigpro.2018.02.031

DO - 10.1016/j.sigpro.2018.02.031

M3 - Article

AN - SCOPUS:85043605701

VL - 149

SP - 49

EP - 61

JO - Signal Processing

JF - Signal Processing

SN - 0165-1684

ER -