Robust initialisation for single-plane 3D CT to 2D fluoroscopy image registration

Masuma Akter, Andrew Lambert, Mark Pickering, Jennie Scarvell, Paul Smith

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

Gradient descent-based automatic image registration algorithms typically fail when the initial misalignment between objects is large. This is a major limitation for routine clinical applications. The registration task is even more difficult for multi-modal images because of the nonlinear relationship between the pixel intensities in the images to be aligned. In this paper, we present a fast and accurate multi-modal image registration algorithm which successfully registers three-dimensional (3D) computed tomography to two-dimensional single-plane fluoroscopy data for large initial displacements between the images. Our experimental results show that the proposed approach can increase the range of initial displacements up to ± 20 mm for all translations and up to ± 20° for rotation while maintaining high precision and small bias in all six 3D rigid body transform parameters.

Original languageEnglish
Pages (from-to)147-171
Number of pages25
JournalComputer Methods in Biomechanics and Biomedical Engineering: Imagine and Visualisation
Volume3
Issue number3
DOIs
Publication statusPublished - 2015
Externally publishedYes

Fingerprint

Fluoroscopy
Image registration
Tomography
Pixels

Cite this

@article{d2831ececdfc4d8aabe7d75745b61020,
title = "Robust initialisation for single-plane 3D CT to 2D fluoroscopy image registration",
abstract = "Gradient descent-based automatic image registration algorithms typically fail when the initial misalignment between objects is large. This is a major limitation for routine clinical applications. The registration task is even more difficult for multi-modal images because of the nonlinear relationship between the pixel intensities in the images to be aligned. In this paper, we present a fast and accurate multi-modal image registration algorithm which successfully registers three-dimensional (3D) computed tomography to two-dimensional single-plane fluoroscopy data for large initial displacements between the images. Our experimental results show that the proposed approach can increase the range of initial displacements up to ± 20 mm for all translations and up to ± 20° for rotation while maintaining high precision and small bias in all six 3D rigid body transform parameters.",
keywords = "image registration, knee kinematics, 3D-2D registration, log-polar transform, nonlinear filter, pixel mapping, 3D–2D registration",
author = "Masuma Akter and Andrew Lambert and Mark Pickering and Jennie Scarvell and Paul Smith",
year = "2015",
doi = "10.1080/21681163.2014.897649",
language = "English",
volume = "3",
pages = "147--171",
journal = "Computer Methods in Biomechanics and Biomedical Engineering: Imaging and Visualization",
issn = "2168-1163",
publisher = "Taylor and Francis Ltd.",
number = "3",

}

Robust initialisation for single-plane 3D CT to 2D fluoroscopy image registration. / Akter, Masuma; Lambert, Andrew; Pickering, Mark; Scarvell, Jennie; Smith, Paul.

In: Computer Methods in Biomechanics and Biomedical Engineering: Imagine and Visualisation, Vol. 3, No. 3, 2015, p. 147-171.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Robust initialisation for single-plane 3D CT to 2D fluoroscopy image registration

AU - Akter, Masuma

AU - Lambert, Andrew

AU - Pickering, Mark

AU - Scarvell, Jennie

AU - Smith, Paul

PY - 2015

Y1 - 2015

N2 - Gradient descent-based automatic image registration algorithms typically fail when the initial misalignment between objects is large. This is a major limitation for routine clinical applications. The registration task is even more difficult for multi-modal images because of the nonlinear relationship between the pixel intensities in the images to be aligned. In this paper, we present a fast and accurate multi-modal image registration algorithm which successfully registers three-dimensional (3D) computed tomography to two-dimensional single-plane fluoroscopy data for large initial displacements between the images. Our experimental results show that the proposed approach can increase the range of initial displacements up to ± 20 mm for all translations and up to ± 20° for rotation while maintaining high precision and small bias in all six 3D rigid body transform parameters.

AB - Gradient descent-based automatic image registration algorithms typically fail when the initial misalignment between objects is large. This is a major limitation for routine clinical applications. The registration task is even more difficult for multi-modal images because of the nonlinear relationship between the pixel intensities in the images to be aligned. In this paper, we present a fast and accurate multi-modal image registration algorithm which successfully registers three-dimensional (3D) computed tomography to two-dimensional single-plane fluoroscopy data for large initial displacements between the images. Our experimental results show that the proposed approach can increase the range of initial displacements up to ± 20 mm for all translations and up to ± 20° for rotation while maintaining high precision and small bias in all six 3D rigid body transform parameters.

KW - image registration

KW - knee kinematics

KW - 3D-2D registration

KW - log-polar transform

KW - nonlinear filter

KW - pixel mapping

KW - 3D–2D registration

UR - http://www.scopus.com/inward/record.url?scp=84981360639&partnerID=8YFLogxK

UR - http://www.mendeley.com/research/robust-initialisation-singleplane-3d-ct-2d-fluoroscopy-image-registration

U2 - 10.1080/21681163.2014.897649

DO - 10.1080/21681163.2014.897649

M3 - Article

VL - 3

SP - 147

EP - 171

JO - Computer Methods in Biomechanics and Biomedical Engineering: Imaging and Visualization

JF - Computer Methods in Biomechanics and Biomedical Engineering: Imaging and Visualization

SN - 2168-1163

IS - 3

ER -