A Statistical Model for Rigid Image Registration Performance: The Influence of Soft-Tissue Deformation as a Confounding Noise Source

Michael D. Ketcha, Tharindu De Silva, Runze Han, Ali Uneri, Sebastian Vogt, Gerhard Kleinszig, Jeffrey H. Siewerdsen

Research output: Contribution to journalArticle

Abstract

Soft-tissue deformation presents a confounding factor to rigid image registration by introducing image content inconsistent with the underlying motion model, presenting non-correspondent structure with potentially high power, and creating local minima that challenge iterative optimization. In this paper, we introduce a model for registration performance that includes deformable soft tissue as a power-law noise distribution within a statistical framework describing the Cramer-Rao lower bound (CRLB) and root-mean-squared error (RMSE) in registration performance. The model incorporates both cross-correlation and gradient-based similarity metrics, and the model was tested in application to 3D-2D (CT-to-radiograph) and 3D-3D (CT-to-CT) image registration. Predictions accurately reflect the trends in registration error as a function of dose (quantum noise), and the choice of similarity metrics for both registration scenarios. Incorporating soft-tissue deformation as a noise source yields important insight on the limits of registration performance with respect to algorithm design and the clinical application or anatomical context. For example, the model quantifies the advantage of gradient-based similarity metrics in 3D-2D registration, identifies the low-dose limits of registration performance, and reveals the conditions for which the registration performance is fundamentally limited by soft-tissue deformation.

Original languageEnglish (US)
Pages (from-to)2016-2027
Number of pages12
JournalIEEE transactions on medical imaging
Volume38
Issue number9
DOIs
StatePublished - Sep 1 2019

Fingerprint

Image registration
Statistical Models
Noise
Tissue
Quantum noise

ASJC Scopus subject areas

  • Software
  • Radiological and Ultrasound Technology
  • Computer Science Applications
  • Electrical and Electronic Engineering

Cite this

A Statistical Model for Rigid Image Registration Performance : The Influence of Soft-Tissue Deformation as a Confounding Noise Source. / Ketcha, Michael D.; De Silva, Tharindu; Han, Runze; Uneri, Ali; Vogt, Sebastian; Kleinszig, Gerhard; Siewerdsen, Jeffrey H.

In: IEEE transactions on medical imaging, Vol. 38, No. 9, 01.09.2019, p. 2016-2027.

Research output: Contribution to journalArticle

Ketcha, Michael D. ; De Silva, Tharindu ; Han, Runze ; Uneri, Ali ; Vogt, Sebastian ; Kleinszig, Gerhard ; Siewerdsen, Jeffrey H. / A Statistical Model for Rigid Image Registration Performance : The Influence of Soft-Tissue Deformation as a Confounding Noise Source. In: IEEE transactions on medical imaging. 2019 ; Vol. 38, No. 9. pp. 2016-2027.
@article{dafd14a2048f405c828ef3ef39bda365,
title = "A Statistical Model for Rigid Image Registration Performance: The Influence of Soft-Tissue Deformation as a Confounding Noise Source",
abstract = "Soft-tissue deformation presents a confounding factor to rigid image registration by introducing image content inconsistent with the underlying motion model, presenting non-correspondent structure with potentially high power, and creating local minima that challenge iterative optimization. In this paper, we introduce a model for registration performance that includes deformable soft tissue as a power-law noise distribution within a statistical framework describing the Cramer-Rao lower bound (CRLB) and root-mean-squared error (RMSE) in registration performance. The model incorporates both cross-correlation and gradient-based similarity metrics, and the model was tested in application to 3D-2D (CT-to-radiograph) and 3D-3D (CT-to-CT) image registration. Predictions accurately reflect the trends in registration error as a function of dose (quantum noise), and the choice of similarity metrics for both registration scenarios. Incorporating soft-tissue deformation as a noise source yields important insight on the limits of registration performance with respect to algorithm design and the clinical application or anatomical context. For example, the model quantifies the advantage of gradient-based similarity metrics in 3D-2D registration, identifies the low-dose limits of registration performance, and reveals the conditions for which the registration performance is fundamentally limited by soft-tissue deformation.",
author = "Ketcha, {Michael D.} and {De Silva}, Tharindu and Runze Han and Ali Uneri and Sebastian Vogt and Gerhard Kleinszig and Siewerdsen, {Jeffrey H.}",
year = "2019",
month = "9",
day = "1",
doi = "10.1109/TMI.2019.2907868",
language = "English (US)",
volume = "38",
pages = "2016--2027",
journal = "IEEE Transactions on Medical Imaging",
issn = "0278-0062",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "9",

}

TY - JOUR

T1 - A Statistical Model for Rigid Image Registration Performance

T2 - The Influence of Soft-Tissue Deformation as a Confounding Noise Source

AU - Ketcha, Michael D.

AU - De Silva, Tharindu

AU - Han, Runze

AU - Uneri, Ali

AU - Vogt, Sebastian

AU - Kleinszig, Gerhard

AU - Siewerdsen, Jeffrey H.

PY - 2019/9/1

Y1 - 2019/9/1

N2 - Soft-tissue deformation presents a confounding factor to rigid image registration by introducing image content inconsistent with the underlying motion model, presenting non-correspondent structure with potentially high power, and creating local minima that challenge iterative optimization. In this paper, we introduce a model for registration performance that includes deformable soft tissue as a power-law noise distribution within a statistical framework describing the Cramer-Rao lower bound (CRLB) and root-mean-squared error (RMSE) in registration performance. The model incorporates both cross-correlation and gradient-based similarity metrics, and the model was tested in application to 3D-2D (CT-to-radiograph) and 3D-3D (CT-to-CT) image registration. Predictions accurately reflect the trends in registration error as a function of dose (quantum noise), and the choice of similarity metrics for both registration scenarios. Incorporating soft-tissue deformation as a noise source yields important insight on the limits of registration performance with respect to algorithm design and the clinical application or anatomical context. For example, the model quantifies the advantage of gradient-based similarity metrics in 3D-2D registration, identifies the low-dose limits of registration performance, and reveals the conditions for which the registration performance is fundamentally limited by soft-tissue deformation.

AB - Soft-tissue deformation presents a confounding factor to rigid image registration by introducing image content inconsistent with the underlying motion model, presenting non-correspondent structure with potentially high power, and creating local minima that challenge iterative optimization. In this paper, we introduce a model for registration performance that includes deformable soft tissue as a power-law noise distribution within a statistical framework describing the Cramer-Rao lower bound (CRLB) and root-mean-squared error (RMSE) in registration performance. The model incorporates both cross-correlation and gradient-based similarity metrics, and the model was tested in application to 3D-2D (CT-to-radiograph) and 3D-3D (CT-to-CT) image registration. Predictions accurately reflect the trends in registration error as a function of dose (quantum noise), and the choice of similarity metrics for both registration scenarios. Incorporating soft-tissue deformation as a noise source yields important insight on the limits of registration performance with respect to algorithm design and the clinical application or anatomical context. For example, the model quantifies the advantage of gradient-based similarity metrics in 3D-2D registration, identifies the low-dose limits of registration performance, and reveals the conditions for which the registration performance is fundamentally limited by soft-tissue deformation.

UR - http://www.scopus.com/inward/record.url?scp=85071785864&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85071785864&partnerID=8YFLogxK

U2 - 10.1109/TMI.2019.2907868

DO - 10.1109/TMI.2019.2907868

M3 - Article

C2 - 30932834

AN - SCOPUS:85071785864

VL - 38

SP - 2016

EP - 2027

JO - IEEE Transactions on Medical Imaging

JF - IEEE Transactions on Medical Imaging

SN - 0278-0062

IS - 9

ER -