Image Restoration with Multiple Hard Constraints on Data-Fidelity to Blurred/Noisy Image Pair

Saori TAKEYAMA; Shunsuke ONO; Itsuo KUMAZAWA

doi:10.1587/transinf.2016PCP0003

Image Restoration with Multiple Hard Constraints on Data-Fidelity to Blurred/Noisy Image Pair

Saori TAKEYAMA, Shunsuke ONO, Itsuo KUMAZAWA

Full Text Views

0

Cite this

Summary :

Existing image deblurring methods with a blurred/noisy image pair take a two-step approach: blur kernel estimation and image restoration. They can achieve better and much more stable blur kernel estimation than single image deblurring methods. On the other hand, in the image restoration step, they do not exploit the information on the noisy image, or they require ad hoc tuning of interdependent parameters. This paper focuses on the image restoration step and proposes a new restoration method of using a blurred/noisy image pair. In our method, the image restoration problem is formulated as a constrained convex optimization problem, where data-fidelity to a blurred image and that to a noisy image is properly taken into account as multiple hard constraints. This offers (i) high quality restoration when the blurred image also contains noise; (ii) robustness to the estimation error of the blur kernel; and (iii) easy parameter setting. We also provide an efficient algorithm for solving our optimization problem based on the so-called alternating direction method of multipliers (ADMM). Experimental results support our claims.

Publication: IEICE TRANSACTIONS on Information Vol.E100-D No.9 pp.1953-1961

Publication Date: 2017/09/01

Publicized: 2017/06/14

Online ISSN: 1745-1361

DOI: 10.1587/transinf.2016PCP0003

Type of Manuscript: Special Section PAPER (Special Section on Picture Coding and Image Media Processing)

Category

Authors

Saori TAKEYAMA
  Department of Information and Communications Engineering at the Tokyo Institute of Technology
Shunsuke ONO
  Tokyo Institute of Technology,Laboratory for Future Interdisciplinary Research of Science and Technology (FIRST)
Itsuo KUMAZAWA
  Tokyo Institute of Technology,Laboratory for Future Interdisciplinary Research of Science and Technology (FIRST)

Keyword

ADMM, deblurring, hard constraints, image restoration, constrained convex optimization

Cite this

Copy

Saori TAKEYAMA, Shunsuke ONO, Itsuo KUMAZAWA, "Image Restoration with Multiple Hard Constraints on Data-Fidelity to Blurred/Noisy Image Pair" in IEICE TRANSACTIONS on Information, vol. E100-D, no. 9, pp. 1953-1961, September 2017, doi: 10.1587/transinf.2016PCP0003.
Abstract: Existing image deblurring methods with a blurred/noisy image pair take a two-step approach: blur kernel estimation and image restoration. They can achieve better and much more stable blur kernel estimation than single image deblurring methods. On the other hand, in the image restoration step, they do not exploit the information on the noisy image, or they require ad hoc tuning of interdependent parameters. This paper focuses on the image restoration step and proposes a new restoration method of using a blurred/noisy image pair. In our method, the image restoration problem is formulated as a constrained convex optimization problem, where data-fidelity to a blurred image and that to a noisy image is properly taken into account as multiple hard constraints. This offers (i) high quality restoration when the blurred image also contains noise; (ii) robustness to the estimation error of the blur kernel; and (iii) easy parameter setting. We also provide an efficient algorithm for solving our optimization problem based on the so-called alternating direction method of multipliers (ADMM). Experimental results support our claims.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2016PCP0003/_p

Copy

@ARTICLE{e100-d_9_1953,
author={Saori TAKEYAMA, Shunsuke ONO, Itsuo KUMAZAWA, },
journal={IEICE TRANSACTIONS on Information},
title={Image Restoration with Multiple Hard Constraints on Data-Fidelity to Blurred/Noisy Image Pair},
year={2017},
volume={E100-D},
number={9},
pages={1953-1961},
abstract={Existing image deblurring methods with a blurred/noisy image pair take a two-step approach: blur kernel estimation and image restoration. They can achieve better and much more stable blur kernel estimation than single image deblurring methods. On the other hand, in the image restoration step, they do not exploit the information on the noisy image, or they require ad hoc tuning of interdependent parameters. This paper focuses on the image restoration step and proposes a new restoration method of using a blurred/noisy image pair. In our method, the image restoration problem is formulated as a constrained convex optimization problem, where data-fidelity to a blurred image and that to a noisy image is properly taken into account as multiple hard constraints. This offers (i) high quality restoration when the blurred image also contains noise; (ii) robustness to the estimation error of the blur kernel; and (iii) easy parameter setting. We also provide an efficient algorithm for solving our optimization problem based on the so-called alternating direction method of multipliers (ADMM). Experimental results support our claims.},
keywords={},
doi={10.1587/transinf.2016PCP0003},
ISSN={1745-1361},
month={September},}

Copy

TY - JOUR
TI - Image Restoration with Multiple Hard Constraints on Data-Fidelity to Blurred/Noisy Image Pair
T2 - IEICE TRANSACTIONS on Information
SP - 1953
EP - 1961
AU - Saori TAKEYAMA
AU - Shunsuke ONO
AU - Itsuo KUMAZAWA
PY - 2017
DO - 10.1587/transinf.2016PCP0003
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E100-D
IS - 9
JA - IEICE TRANSACTIONS on Information
Y1 - September 2017
AB - Existing image deblurring methods with a blurred/noisy image pair take a two-step approach: blur kernel estimation and image restoration. They can achieve better and much more stable blur kernel estimation than single image deblurring methods. On the other hand, in the image restoration step, they do not exploit the information on the noisy image, or they require ad hoc tuning of interdependent parameters. This paper focuses on the image restoration step and proposes a new restoration method of using a blurred/noisy image pair. In our method, the image restoration problem is formulated as a constrained convex optimization problem, where data-fidelity to a blurred image and that to a noisy image is properly taken into account as multiple hard constraints. This offers (i) high quality restoration when the blurred image also contains noise; (ii) robustness to the estimation error of the blur kernel; and (iii) easy parameter setting. We also provide an efficient algorithm for solving our optimization problem based on the so-called alternating direction method of multipliers (ADMM). Experimental results support our claims.
ER -