Imaging spatially varying biomechanical properties with neural networks

Cameron Hoerig, Wendy Reyes, Léo Fabre, Jamshid Ghaboussi, Michael Insana

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Elastography comprises a set of modalities that image the biomechanical properties of soft tissues for disease detection and diagnosis. Quasi-static ultrasound elastography, in particular, tracks sub-surface displacements resulting from an applied surface force. The local displacement information and measured surface loads may be used to compute a parametric summary of biomechanical properties; however, the inverse problem is under-determined, limiting most techniques to estimating a single linear-elastic parameter. We previously described a new method to develop mechanical models using a combination of computational mechanics and machine learning that circumvents the limitations associated with the inverse problem. The Autoprogressive method weaves together finite element analysis and artificial neural networks (ANNs) to develop empirical models of mechanical behavior using only measured force-displacement data. We are extending that work by incorporating spatial information with the material properties. Previously, the ANNs accepted only a strain vector input and computed the corresponding stress, meaning any spatial information was encoded in the finite element mesh. Now, using a pair of ANNs working in tandem with spatial coordinates included as part of the input, these new Cartesian ANNs are able to learn the spatially varying mechanical behavior of complex media. We show that a single Cartesian ANN is able to describe the same mechanical behavior of an object that previously required at least two ANNs. Furthermore, we show the new ANNs can learn complex material property distributions and reconstruct images of the Young's modulus distribution, not merely classify, filter, or otherwise process an existing image. For the first time, we present results using Cartesian neural networks within the Autoprogressive Method to form elastic modulus images.

Original languageEnglish (US)
Title of host publicationMedical Imaging 2017
Subtitle of host publicationUltrasonic Imaging and Tomography
EditorsNeb Duric, Neb Duric, Brecht Heyde
PublisherSPIE
ISBN (Electronic)9781510607231
DOIs
StatePublished - Jan 1 2017
EventMedical Imaging 2017: Ultrasonic Imaging and Tomography - Orlando, United States
Duration: Feb 15 2017Feb 16 2017

Publication series

NameProgress in Biomedical Optics and Imaging - Proceedings of SPIE
Volume10139
ISSN (Print)1605-7422

Other

OtherMedical Imaging 2017: Ultrasonic Imaging and Tomography
CountryUnited States
CityOrlando
Period2/15/172/16/17

Fingerprint

Elasticity Imaging Techniques
Elastic Modulus
Neural networks
Imaging techniques
Finite Element Analysis
Mechanics
Inverse problems
Materials properties
modulus of elasticity
computational mechanics
Elastic moduli
Computational mechanics
machine learning
distribution (property)
learning
Learning systems
mesh
estimating
Ultrasonics
Tissue

Keywords

  • Constitutive Modeling
  • Elasticity Imaging
  • Finite Element Analysis
  • Machine Learning
  • Ultrasound

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Biomaterials
  • Atomic and Molecular Physics, and Optics
  • Radiology Nuclear Medicine and imaging

Cite this

Hoerig, C., Reyes, W., Fabre, L., Ghaboussi, J., & Insana, M. (2017). Imaging spatially varying biomechanical properties with neural networks. In N. Duric, N. Duric, & B. Heyde (Eds.), Medical Imaging 2017: Ultrasonic Imaging and Tomography [1013905] (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 10139). SPIE. https://doi.org/10.1117/12.2254331

Imaging spatially varying biomechanical properties with neural networks. / Hoerig, Cameron; Reyes, Wendy; Fabre, Léo; Ghaboussi, Jamshid; Insana, Michael.

Medical Imaging 2017: Ultrasonic Imaging and Tomography. ed. / Neb Duric; Neb Duric; Brecht Heyde. SPIE, 2017. 1013905 (Progress in Biomedical Optics and Imaging - Proceedings of SPIE; Vol. 10139).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Hoerig, C, Reyes, W, Fabre, L, Ghaboussi, J & Insana, M 2017, Imaging spatially varying biomechanical properties with neural networks. in N Duric, N Duric & B Heyde (eds), Medical Imaging 2017: Ultrasonic Imaging and Tomography., 1013905, Progress in Biomedical Optics and Imaging - Proceedings of SPIE, vol. 10139, SPIE, Medical Imaging 2017: Ultrasonic Imaging and Tomography, Orlando, United States, 2/15/17. https://doi.org/10.1117/12.2254331
Hoerig C, Reyes W, Fabre L, Ghaboussi J, Insana M. Imaging spatially varying biomechanical properties with neural networks. In Duric N, Duric N, Heyde B, editors, Medical Imaging 2017: Ultrasonic Imaging and Tomography. SPIE. 2017. 1013905. (Progress in Biomedical Optics and Imaging - Proceedings of SPIE). https://doi.org/10.1117/12.2254331
Hoerig, Cameron ; Reyes, Wendy ; Fabre, Léo ; Ghaboussi, Jamshid ; Insana, Michael. / Imaging spatially varying biomechanical properties with neural networks. Medical Imaging 2017: Ultrasonic Imaging and Tomography. editor / Neb Duric ; Neb Duric ; Brecht Heyde. SPIE, 2017. (Progress in Biomedical Optics and Imaging - Proceedings of SPIE).
@inproceedings{128e2141a1b14eb78c9537e2fc6ffff8,
title = "Imaging spatially varying biomechanical properties with neural networks",
abstract = "Elastography comprises a set of modalities that image the biomechanical properties of soft tissues for disease detection and diagnosis. Quasi-static ultrasound elastography, in particular, tracks sub-surface displacements resulting from an applied surface force. The local displacement information and measured surface loads may be used to compute a parametric summary of biomechanical properties; however, the inverse problem is under-determined, limiting most techniques to estimating a single linear-elastic parameter. We previously described a new method to develop mechanical models using a combination of computational mechanics and machine learning that circumvents the limitations associated with the inverse problem. The Autoprogressive method weaves together finite element analysis and artificial neural networks (ANNs) to develop empirical models of mechanical behavior using only measured force-displacement data. We are extending that work by incorporating spatial information with the material properties. Previously, the ANNs accepted only a strain vector input and computed the corresponding stress, meaning any spatial information was encoded in the finite element mesh. Now, using a pair of ANNs working in tandem with spatial coordinates included as part of the input, these new Cartesian ANNs are able to learn the spatially varying mechanical behavior of complex media. We show that a single Cartesian ANN is able to describe the same mechanical behavior of an object that previously required at least two ANNs. Furthermore, we show the new ANNs can learn complex material property distributions and reconstruct images of the Young's modulus distribution, not merely classify, filter, or otherwise process an existing image. For the first time, we present results using Cartesian neural networks within the Autoprogressive Method to form elastic modulus images.",
keywords = "Constitutive Modeling, Elasticity Imaging, Finite Element Analysis, Machine Learning, Ultrasound",
author = "Cameron Hoerig and Wendy Reyes and L{\'e}o Fabre and Jamshid Ghaboussi and Michael Insana",
year = "2017",
month = "1",
day = "1",
doi = "10.1117/12.2254331",
language = "English (US)",
series = "Progress in Biomedical Optics and Imaging - Proceedings of SPIE",
publisher = "SPIE",
editor = "Neb Duric and Neb Duric and Brecht Heyde",
booktitle = "Medical Imaging 2017",

}

TY - GEN

T1 - Imaging spatially varying biomechanical properties with neural networks

AU - Hoerig, Cameron

AU - Reyes, Wendy

AU - Fabre, Léo

AU - Ghaboussi, Jamshid

AU - Insana, Michael

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Elastography comprises a set of modalities that image the biomechanical properties of soft tissues for disease detection and diagnosis. Quasi-static ultrasound elastography, in particular, tracks sub-surface displacements resulting from an applied surface force. The local displacement information and measured surface loads may be used to compute a parametric summary of biomechanical properties; however, the inverse problem is under-determined, limiting most techniques to estimating a single linear-elastic parameter. We previously described a new method to develop mechanical models using a combination of computational mechanics and machine learning that circumvents the limitations associated with the inverse problem. The Autoprogressive method weaves together finite element analysis and artificial neural networks (ANNs) to develop empirical models of mechanical behavior using only measured force-displacement data. We are extending that work by incorporating spatial information with the material properties. Previously, the ANNs accepted only a strain vector input and computed the corresponding stress, meaning any spatial information was encoded in the finite element mesh. Now, using a pair of ANNs working in tandem with spatial coordinates included as part of the input, these new Cartesian ANNs are able to learn the spatially varying mechanical behavior of complex media. We show that a single Cartesian ANN is able to describe the same mechanical behavior of an object that previously required at least two ANNs. Furthermore, we show the new ANNs can learn complex material property distributions and reconstruct images of the Young's modulus distribution, not merely classify, filter, or otherwise process an existing image. For the first time, we present results using Cartesian neural networks within the Autoprogressive Method to form elastic modulus images.

AB - Elastography comprises a set of modalities that image the biomechanical properties of soft tissues for disease detection and diagnosis. Quasi-static ultrasound elastography, in particular, tracks sub-surface displacements resulting from an applied surface force. The local displacement information and measured surface loads may be used to compute a parametric summary of biomechanical properties; however, the inverse problem is under-determined, limiting most techniques to estimating a single linear-elastic parameter. We previously described a new method to develop mechanical models using a combination of computational mechanics and machine learning that circumvents the limitations associated with the inverse problem. The Autoprogressive method weaves together finite element analysis and artificial neural networks (ANNs) to develop empirical models of mechanical behavior using only measured force-displacement data. We are extending that work by incorporating spatial information with the material properties. Previously, the ANNs accepted only a strain vector input and computed the corresponding stress, meaning any spatial information was encoded in the finite element mesh. Now, using a pair of ANNs working in tandem with spatial coordinates included as part of the input, these new Cartesian ANNs are able to learn the spatially varying mechanical behavior of complex media. We show that a single Cartesian ANN is able to describe the same mechanical behavior of an object that previously required at least two ANNs. Furthermore, we show the new ANNs can learn complex material property distributions and reconstruct images of the Young's modulus distribution, not merely classify, filter, or otherwise process an existing image. For the first time, we present results using Cartesian neural networks within the Autoprogressive Method to form elastic modulus images.

KW - Constitutive Modeling

KW - Elasticity Imaging

KW - Finite Element Analysis

KW - Machine Learning

KW - Ultrasound

UR - http://www.scopus.com/inward/record.url?scp=85020767407&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85020767407&partnerID=8YFLogxK

U2 - 10.1117/12.2254331

DO - 10.1117/12.2254331

M3 - Conference contribution

AN - SCOPUS:85020767407

T3 - Progress in Biomedical Optics and Imaging - Proceedings of SPIE

BT - Medical Imaging 2017

A2 - Duric, Neb

A2 - Duric, Neb

A2 - Heyde, Brecht

PB - SPIE

ER -