2013
DOI: 10.1109/tip.2013.2268972
|View full text |Cite
|
Sign up to set email alerts
|

Blur Invariant Translational Image Registration for $N$-fold Symmetric Blurs

Abstract: In this paper, we propose a new registration method designed particularly for registering differently blurred images. Such a task cannot be successfully resolved by traditional approaches. Our method is inspired by traditional phase correlation, which is now applied to certain blur-invariant descriptors instead of the original images. This method works for unknown blurs assuming the blurring PSF exhibits an N-fold rotational symmetry. It does not require any landmarks. We have experimentally proven its good pe… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
2
1

Citation Types

0
9
0

Year Published

2014
2014
2019
2019

Publication Types

Select...
4
3
3

Relationship

3
7

Authors

Journals

citations
Cited by 21 publications
(9 citation statements)
references
References 34 publications
0
9
0
Order By: Relevance
“…Under these premises, the Inverse Fourier Transform results in a single peak, which the location of the peak corresponds to the motion vectors between the two frames. For heavily blurred movies, we utilized an improved version of the original phase correlation method [ 49 ], which is more robust in registering blurred images. During the motion compensation step, maximum overlap among the frames was calculated and the point of view was fixed by shifting frame centers according to estimated motion vectors in the previous step.…”
Section: Methodsmentioning
confidence: 99%
“…Under these premises, the Inverse Fourier Transform results in a single peak, which the location of the peak corresponds to the motion vectors between the two frames. For heavily blurred movies, we utilized an improved version of the original phase correlation method [ 49 ], which is more robust in registering blurred images. During the motion compensation step, maximum overlap among the frames was calculated and the point of view was fixed by shifting frame centers according to estimated motion vectors in the previous step.…”
Section: Methodsmentioning
confidence: 99%
“…The shift error is defined as the Euclidean distance between the true shift and the estimated shift. The performance is quantified according to the criterion proposed in [16], that is, by counting the number of misregistrations, where a "misregistration" occurs whenever exceeds a specified threshold . We set pixels.…”
Section: Methodsmentioning
confidence: 99%
“…Global methods do not search for particular landmarks in the images but rather try to estimate the between-image transformation directly. Most blur-invariant global methods were motivated by traditional phase correlation [84], where the blur-insensitivity was achieved by modifications of the cross-power spectrum [41], [46], [47], [85]. Global methods are fast and easy to implement, but their limiting requirements -simple betweenframe distortion (most of them allow only translation and/or rotation), the need for a large overlap of the images and the assumption of the uniform blur -might be a drawback in more complicated situations.…”
Section: Registration Of Blurred Imagesmentioning
confidence: 99%