Image restoration with multiple DirLOTs

Natsuki Aizawa, Shogo Muramatsu, Masahiro Yukawa

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

A directional lapped orthogonal transform (DirLOT) is an orthonormal transform of which basis is allowed to be anisotropic with the symmetric, real-valued and compact-support property. Due to its directional property, DirLOT is superior to the existing separable transforms such as DCT and DWT in expressing diagonal edges and textures. The goal of this paper is to enhance the ability of DirLOT further. To achieve this goal, we propose a novel image restoration technique using multiple DirLOTs. This paper generalizes an image denoising technique in [1], and expands the application of multiple DirLOTs by introducing linear degradation operator P. The idea is to use multiple DirLOTs to construct a redundant dictionary. More precisely, the redundant dictionary is constructed as a union of symmetric orthonormal discrete wavelet transforms generated by DirLOTs. To select atoms fitting a target image from the dictionary, we formulate an image restoration problem as an ℓ1- regularized least square problem, which can efficiently be solved by the iterativeshrinkage/thresholding algorithm (ISTA). The proposed technique is beneficial in expressing multiple directions of edges/textures. Simulation results show that the proposed technique significantly outperforms the nonsubsampled Haar wavelet transform for deblurring, super-resolution, and inpainting.

Original languageEnglish
Pages (from-to)1954-1961
Number of pages8
JournalIEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences
VolumeE96-A
Issue number10
DOIs
Publication statusPublished - 2013 Oct

Fingerprint

Image Restoration
Glossaries
Image reconstruction
Transform
Textures
Orthonormal
Image denoising
Discrete wavelet transforms
Wavelet Transform
Texture
Wavelet transforms
Deblurring
Inpainting
Haar Wavelet
Image Denoising
Super-resolution
Least Squares Problem
Compact Support
Degradation
Atoms

Keywords

  • Deblurring
  • DirLOT
  • Inpainting
  • Iterativeshrinkage/ thresholding algorithm (ISTA)
  • Multi-directional dictionary
  • Super-resolution

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Computer Graphics and Computer-Aided Design
  • Applied Mathematics
  • Signal Processing

Cite this

Image restoration with multiple DirLOTs. / Aizawa, Natsuki; Muramatsu, Shogo; Yukawa, Masahiro.

In: IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences, Vol. E96-A, No. 10, 10.2013, p. 1954-1961.

Research output: Contribution to journalArticle

@article{090367c017be4a2f8931abab3bcf63e8,
title = "Image restoration with multiple DirLOTs",
abstract = "A directional lapped orthogonal transform (DirLOT) is an orthonormal transform of which basis is allowed to be anisotropic with the symmetric, real-valued and compact-support property. Due to its directional property, DirLOT is superior to the existing separable transforms such as DCT and DWT in expressing diagonal edges and textures. The goal of this paper is to enhance the ability of DirLOT further. To achieve this goal, we propose a novel image restoration technique using multiple DirLOTs. This paper generalizes an image denoising technique in [1], and expands the application of multiple DirLOTs by introducing linear degradation operator P. The idea is to use multiple DirLOTs to construct a redundant dictionary. More precisely, the redundant dictionary is constructed as a union of symmetric orthonormal discrete wavelet transforms generated by DirLOTs. To select atoms fitting a target image from the dictionary, we formulate an image restoration problem as an ℓ1- regularized least square problem, which can efficiently be solved by the iterativeshrinkage/thresholding algorithm (ISTA). The proposed technique is beneficial in expressing multiple directions of edges/textures. Simulation results show that the proposed technique significantly outperforms the nonsubsampled Haar wavelet transform for deblurring, super-resolution, and inpainting.",
keywords = "Deblurring, DirLOT, Inpainting, Iterativeshrinkage/ thresholding algorithm (ISTA), Multi-directional dictionary, Super-resolution",
author = "Natsuki Aizawa and Shogo Muramatsu and Masahiro Yukawa",
year = "2013",
month = "10",
doi = "10.1587/transfun.E96.A.1954",
language = "English",
volume = "E96-A",
pages = "1954--1961",
journal = "IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences",
issn = "0916-8508",
publisher = "Maruzen Co., Ltd/Maruzen Kabushikikaisha",
number = "10",

}

TY - JOUR

T1 - Image restoration with multiple DirLOTs

AU - Aizawa, Natsuki

AU - Muramatsu, Shogo

AU - Yukawa, Masahiro

PY - 2013/10

Y1 - 2013/10

N2 - A directional lapped orthogonal transform (DirLOT) is an orthonormal transform of which basis is allowed to be anisotropic with the symmetric, real-valued and compact-support property. Due to its directional property, DirLOT is superior to the existing separable transforms such as DCT and DWT in expressing diagonal edges and textures. The goal of this paper is to enhance the ability of DirLOT further. To achieve this goal, we propose a novel image restoration technique using multiple DirLOTs. This paper generalizes an image denoising technique in [1], and expands the application of multiple DirLOTs by introducing linear degradation operator P. The idea is to use multiple DirLOTs to construct a redundant dictionary. More precisely, the redundant dictionary is constructed as a union of symmetric orthonormal discrete wavelet transforms generated by DirLOTs. To select atoms fitting a target image from the dictionary, we formulate an image restoration problem as an ℓ1- regularized least square problem, which can efficiently be solved by the iterativeshrinkage/thresholding algorithm (ISTA). The proposed technique is beneficial in expressing multiple directions of edges/textures. Simulation results show that the proposed technique significantly outperforms the nonsubsampled Haar wavelet transform for deblurring, super-resolution, and inpainting.

AB - A directional lapped orthogonal transform (DirLOT) is an orthonormal transform of which basis is allowed to be anisotropic with the symmetric, real-valued and compact-support property. Due to its directional property, DirLOT is superior to the existing separable transforms such as DCT and DWT in expressing diagonal edges and textures. The goal of this paper is to enhance the ability of DirLOT further. To achieve this goal, we propose a novel image restoration technique using multiple DirLOTs. This paper generalizes an image denoising technique in [1], and expands the application of multiple DirLOTs by introducing linear degradation operator P. The idea is to use multiple DirLOTs to construct a redundant dictionary. More precisely, the redundant dictionary is constructed as a union of symmetric orthonormal discrete wavelet transforms generated by DirLOTs. To select atoms fitting a target image from the dictionary, we formulate an image restoration problem as an ℓ1- regularized least square problem, which can efficiently be solved by the iterativeshrinkage/thresholding algorithm (ISTA). The proposed technique is beneficial in expressing multiple directions of edges/textures. Simulation results show that the proposed technique significantly outperforms the nonsubsampled Haar wavelet transform for deblurring, super-resolution, and inpainting.

KW - Deblurring

KW - DirLOT

KW - Inpainting

KW - Iterativeshrinkage/ thresholding algorithm (ISTA)

KW - Multi-directional dictionary

KW - Super-resolution

UR - http://www.scopus.com/inward/record.url?scp=84885066609&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84885066609&partnerID=8YFLogxK

U2 - 10.1587/transfun.E96.A.1954

DO - 10.1587/transfun.E96.A.1954

M3 - Article

AN - SCOPUS:84885066609

VL - E96-A

SP - 1954

EP - 1961

JO - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

JF - IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

SN - 0916-8508

IS - 10

ER -