Standard

Stochastic optimization with randomized smoothing for image registration. / Sun, Wei; Poot, Dirk H J; Smal, Ihor; Yang, Xuan; Niessen, Wiro J.; Klein, Stefan.

In: Medical Image Analysis, Vol. 35, 2017, p. 146-158.

Research output: Contribution to journalArticleScientificpeer-review

Harvard

APA

Vancouver

Author

Sun, Wei ; Poot, Dirk H J ; Smal, Ihor ; Yang, Xuan ; Niessen, Wiro J. ; Klein, Stefan. / Stochastic optimization with randomized smoothing for image registration. In: Medical Image Analysis. 2017 ; Vol. 35. pp. 146-158.

BibTeX

@article{54341eb842824d17aa74ff41bbe65ff2,
title = "Stochastic optimization with randomized smoothing for image registration",
abstract = "Image registration is typically formulated as an optimization process, which aims to find the optimal transformation parameters of a given transformation model by minimizing a cost function. Local minima may exist in the optimization landscape, which could hamper the optimization process. To eliminate local minima, smoothing the cost function would be desirable. In this paper, we investigate the use of a randomized smoothing (RS) technique for stochastic gradient descent (SGD) optimization, to effectively smooth the cost function. In this approach, Gaussian noise is added to the transformation parameters prior to computing the cost function gradient in each iteration of the SGD optimizer. The approach is suitable for both rigid and nonrigid registrations. Experiments on synthetic images, cell images, public CT lung data, and public MR brain data demonstrate the effectiveness of the novel RS technique in terms of registration accuracy and robustness.",
keywords = "Image registration, Local minima, Randomized smoothing, Stochastic gradient descent",
author = "Wei Sun and Poot, {Dirk H J} and Ihor Smal and Xuan Yang and Niessen, {Wiro J.} and Stefan Klein",
note = "Accepted Author Manuscript",
year = "2017",
doi = "10.1016/j.media.2016.07.003",
language = "English",
volume = "35",
pages = "146--158",
journal = "Medical Image Analysis",
issn = "1361-8415",
publisher = "Elsevier",

}

RIS

TY - JOUR

T1 - Stochastic optimization with randomized smoothing for image registration

AU - Sun, Wei

AU - Poot, Dirk H J

AU - Smal, Ihor

AU - Yang, Xuan

AU - Niessen, Wiro J.

AU - Klein, Stefan

N1 - Accepted Author Manuscript

PY - 2017

Y1 - 2017

N2 - Image registration is typically formulated as an optimization process, which aims to find the optimal transformation parameters of a given transformation model by minimizing a cost function. Local minima may exist in the optimization landscape, which could hamper the optimization process. To eliminate local minima, smoothing the cost function would be desirable. In this paper, we investigate the use of a randomized smoothing (RS) technique for stochastic gradient descent (SGD) optimization, to effectively smooth the cost function. In this approach, Gaussian noise is added to the transformation parameters prior to computing the cost function gradient in each iteration of the SGD optimizer. The approach is suitable for both rigid and nonrigid registrations. Experiments on synthetic images, cell images, public CT lung data, and public MR brain data demonstrate the effectiveness of the novel RS technique in terms of registration accuracy and robustness.

AB - Image registration is typically formulated as an optimization process, which aims to find the optimal transformation parameters of a given transformation model by minimizing a cost function. Local minima may exist in the optimization landscape, which could hamper the optimization process. To eliminate local minima, smoothing the cost function would be desirable. In this paper, we investigate the use of a randomized smoothing (RS) technique for stochastic gradient descent (SGD) optimization, to effectively smooth the cost function. In this approach, Gaussian noise is added to the transformation parameters prior to computing the cost function gradient in each iteration of the SGD optimizer. The approach is suitable for both rigid and nonrigid registrations. Experiments on synthetic images, cell images, public CT lung data, and public MR brain data demonstrate the effectiveness of the novel RS technique in terms of registration accuracy and robustness.

KW - Image registration

KW - Local minima

KW - Randomized smoothing

KW - Stochastic gradient descent

UR - http://resolver.tudelft.nl/uuid:54341eb8-4282-4d17-aa74-ff41bbe65ff2

U2 - 10.1016/j.media.2016.07.003

DO - 10.1016/j.media.2016.07.003

M3 - Article

VL - 35

SP - 146

EP - 158

JO - Medical Image Analysis

T2 - Medical Image Analysis

JF - Medical Image Analysis

SN - 1361-8415

ER -

ID: 7141337