Singular value decomposition based fusion for super-resolution image reconstruction

H Haidawati Binti Mohamad Nasir, Vladimir Stankovic, Stephen Marshall

Research output: Contribution to journalArticle

20 Citations (Scopus)

Abstract

In this paper, we address a super-resolution problem of generating a high-resolution image from low-resolution images. The proposed super-resolution method consists of three steps: image registration, singular value decomposition (SVD)-based image fusion and interpolation. The contribution of this work is twofold. First we customize an image registration approach using Scale Invariant Feature Transform (SIFT), Belief Propagation and Random Sampling Consensus (RANSAC) for super-resolution. Second, we propose SVD-based fusion to integrate the important features from the low-resolution images. The proposed image registration and fusion steps effectively maintain the important features and greatly improve the super-resolution results. Results, for a variety of image examples, show that the proposed method successfully generates high-resolution images from low-resolution images.
LanguageEnglish
Pages180–191
Number of pages12
JournalSignal Processing: Image Communication
Volume27
Issue number2
DOIs
Publication statusPublished - Feb 2012

Fingerprint

Singular value decomposition
Image resolution
Image reconstruction
Image registration
Image fusion
Optical resolving power
Interpolation
Mathematical transformations
Sampling

Keywords

  • super-resolution
  • image fusion
  • high-resolution image
  • image resolution

Cite this

@article{eb37591964aa48e79fe5bb5f701ad08a,
title = "Singular value decomposition based fusion for super-resolution image reconstruction",
abstract = "In this paper, we address a super-resolution problem of generating a high-resolution image from low-resolution images. The proposed super-resolution method consists of three steps: image registration, singular value decomposition (SVD)-based image fusion and interpolation. The contribution of this work is twofold. First we customize an image registration approach using Scale Invariant Feature Transform (SIFT), Belief Propagation and Random Sampling Consensus (RANSAC) for super-resolution. Second, we propose SVD-based fusion to integrate the important features from the low-resolution images. The proposed image registration and fusion steps effectively maintain the important features and greatly improve the super-resolution results. Results, for a variety of image examples, show that the proposed method successfully generates high-resolution images from low-resolution images.",
keywords = "super-resolution, image fusion, high-resolution image, image resolution",
author = "{Haidawati Binti Mohamad Nasir}, H and Vladimir Stankovic and Stephen Marshall",
year = "2012",
month = "2",
doi = "10.1016/j.image.2011.12.002",
language = "English",
volume = "27",
pages = "180–191",
journal = "Signal Processing: Image Communication",
issn = "0923-5965",
number = "2",

}

Singular value decomposition based fusion for super-resolution image reconstruction. / Haidawati Binti Mohamad Nasir, H; Stankovic, Vladimir; Marshall, Stephen.

In: Signal Processing: Image Communication, Vol. 27, No. 2, 02.2012, p. 180–191.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Singular value decomposition based fusion for super-resolution image reconstruction

AU - Haidawati Binti Mohamad Nasir, H

AU - Stankovic, Vladimir

AU - Marshall, Stephen

PY - 2012/2

Y1 - 2012/2

N2 - In this paper, we address a super-resolution problem of generating a high-resolution image from low-resolution images. The proposed super-resolution method consists of three steps: image registration, singular value decomposition (SVD)-based image fusion and interpolation. The contribution of this work is twofold. First we customize an image registration approach using Scale Invariant Feature Transform (SIFT), Belief Propagation and Random Sampling Consensus (RANSAC) for super-resolution. Second, we propose SVD-based fusion to integrate the important features from the low-resolution images. The proposed image registration and fusion steps effectively maintain the important features and greatly improve the super-resolution results. Results, for a variety of image examples, show that the proposed method successfully generates high-resolution images from low-resolution images.

AB - In this paper, we address a super-resolution problem of generating a high-resolution image from low-resolution images. The proposed super-resolution method consists of three steps: image registration, singular value decomposition (SVD)-based image fusion and interpolation. The contribution of this work is twofold. First we customize an image registration approach using Scale Invariant Feature Transform (SIFT), Belief Propagation and Random Sampling Consensus (RANSAC) for super-resolution. Second, we propose SVD-based fusion to integrate the important features from the low-resolution images. The proposed image registration and fusion steps effectively maintain the important features and greatly improve the super-resolution results. Results, for a variety of image examples, show that the proposed method successfully generates high-resolution images from low-resolution images.

KW - super-resolution

KW - image fusion

KW - high-resolution image

KW - image resolution

UR - http://www.scopus.com/inward/record.url?scp=84856389921&partnerID=8YFLogxK

U2 - 10.1016/j.image.2011.12.002

DO - 10.1016/j.image.2011.12.002

M3 - Article

VL - 27

SP - 180

EP - 191

JO - Signal Processing: Image Communication

T2 - Signal Processing: Image Communication

JF - Signal Processing: Image Communication

SN - 0923-5965

IS - 2

ER -