Regularization Super-Resolution with Inaccurate Image Registration

Ju LIU; Hua YAN; Jian-de SUN

doi:10.1587/transinf.E92.D.59

Regularization Super-Resolution with Inaccurate Image Registration

Ju LIU, Hua YAN, Jian-de SUN

Full Text Views

0

Cite this

Summary :

Considering the inaccuracy of image registration, we propose a new regularization restoration algorithm to solve the ill-posed super-resolution (SR) problem. Registration error is used to obtain cross-channel error information caused by inaccurate image registration. The registration error is considered as the noise mean added into the within-channel observation noise which is known as Additive White Gaussian Noise (AWGN). Based on this consideration, two constraints are regulated pixel by pixel within the framework of Miller's regularization. Regularization parameters connect the two constraints to construct a cost function. The regularization parameters are estimated adaptively in each pixel in terms of the registration error and in each observation channel in terms of the AWGN. In the iterative implementation of the proposed algorithm, sub-sampling operation and sampling aliasing in the detector model are dealt with respectively to make the restored HR image approach the original one further. The transpose of the sub-sampling operation is implemented by nearest interpolation. Simulations show that the proposed regularization algorithm can restore HR images with much sharper edges and greater SNR improvement.

Publication: IEICE TRANSACTIONS on Information Vol.E92-D No.1 pp.59-68

Publication Date: 2009/01/01

Publicized

Online ISSN: 1745-1361

DOI: 10.1587/transinf.E92.D.59

Type of Manuscript: PAPER

Category: Image Processing and Video Processing

Cite this

Copy

Ju LIU, Hua YAN, Jian-de SUN, "Regularization Super-Resolution with Inaccurate Image Registration" in IEICE TRANSACTIONS on Information, vol. E92-D, no. 1, pp. 59-68, January 2009, doi: 10.1587/transinf.E92.D.59.
Abstract: Considering the inaccuracy of image registration, we propose a new regularization restoration algorithm to solve the ill-posed super-resolution (SR) problem. Registration error is used to obtain cross-channel error information caused by inaccurate image registration. The registration error is considered as the noise mean added into the within-channel observation noise which is known as Additive White Gaussian Noise (AWGN). Based on this consideration, two constraints are regulated pixel by pixel within the framework of Miller's regularization. Regularization parameters connect the two constraints to construct a cost function. The regularization parameters are estimated adaptively in each pixel in terms of the registration error and in each observation channel in terms of the AWGN. In the iterative implementation of the proposed algorithm, sub-sampling operation and sampling aliasing in the detector model are dealt with respectively to make the restored HR image approach the original one further. The transpose of the sub-sampling operation is implemented by nearest interpolation. Simulations show that the proposed regularization algorithm can restore HR images with much sharper edges and greater SNR improvement.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.E92.D.59/_p

Copy

@ARTICLE{e92-d_1_59,
author={Ju LIU, Hua YAN, Jian-de SUN, },
journal={IEICE TRANSACTIONS on Information},
title={Regularization Super-Resolution with Inaccurate Image Registration},
year={2009},
volume={E92-D},
number={1},
pages={59-68},
abstract={Considering the inaccuracy of image registration, we propose a new regularization restoration algorithm to solve the ill-posed super-resolution (SR) problem. Registration error is used to obtain cross-channel error information caused by inaccurate image registration. The registration error is considered as the noise mean added into the within-channel observation noise which is known as Additive White Gaussian Noise (AWGN). Based on this consideration, two constraints are regulated pixel by pixel within the framework of Miller's regularization. Regularization parameters connect the two constraints to construct a cost function. The regularization parameters are estimated adaptively in each pixel in terms of the registration error and in each observation channel in terms of the AWGN. In the iterative implementation of the proposed algorithm, sub-sampling operation and sampling aliasing in the detector model are dealt with respectively to make the restored HR image approach the original one further. The transpose of the sub-sampling operation is implemented by nearest interpolation. Simulations show that the proposed regularization algorithm can restore HR images with much sharper edges and greater SNR improvement.},
keywords={},
doi={10.1587/transinf.E92.D.59},
ISSN={1745-1361},
month={January},}

Copy

TY - JOUR
TI - Regularization Super-Resolution with Inaccurate Image Registration
T2 - IEICE TRANSACTIONS on Information
SP - 59
EP - 68
AU - Ju LIU
AU - Hua YAN
AU - Jian-de SUN
PY - 2009
DO - 10.1587/transinf.E92.D.59
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E92-D
IS - 1
JA - IEICE TRANSACTIONS on Information
Y1 - January 2009
AB - Considering the inaccuracy of image registration, we propose a new regularization restoration algorithm to solve the ill-posed super-resolution (SR) problem. Registration error is used to obtain cross-channel error information caused by inaccurate image registration. The registration error is considered as the noise mean added into the within-channel observation noise which is known as Additive White Gaussian Noise (AWGN). Based on this consideration, two constraints are regulated pixel by pixel within the framework of Miller's regularization. Regularization parameters connect the two constraints to construct a cost function. The regularization parameters are estimated adaptively in each pixel in terms of the registration error and in each observation channel in terms of the AWGN. In the iterative implementation of the proposed algorithm, sub-sampling operation and sampling aliasing in the detector model are dealt with respectively to make the restored HR image approach the original one further. The transpose of the sub-sampling operation is implemented by nearest interpolation. Simulations show that the proposed regularization algorithm can restore HR images with much sharper edges and greater SNR improvement.
ER -