Data-driven elasticity imaging using cartesian neural network constitutive models and the autoprogressive method

Cameron Hoerig, Jamshid Ghaboussi, Michael Insana

Research output: Contribution to journalArticle

Abstract

Quasi-static elasticity imaging techniques rely on model-based mathematical inverse methods to estimate mechanical parameters from force-displacement measurements. These techniques introduce simplifying assumptions that preclude exploration of unknown mechanical properties with potential diagnostic value. We previously reported a data-driven approach to elasticity imaging using artificial neural networks (NNs) that circumvents limitations associated with model-based inverse methods. NN constitutive models can learn stress-strain behavior from force-displacement measurements using the autoprogressive (AutoP) method without prior assumptions of the underlying constitutive model. However, information about internal structure was required. We invented Cartesian NN constitutive models (CaNNCMs) that learn the spatial variations of material properties. We are presenting the first implementation of CaNNCMs trained with AutoP to develop data-driven models of 2-D linear-elastic materials. Both simulated and experimental force-displacement data were used as input to AutoP to show that CaNNCMs are able to model both continuous and discrete material property distributions with no prior information of internal object structure. Furthermore, we demonstrate that CaNNCMs are robust to measurement noise and can reconstruct reasonably accurate Young's modulus images from a sparse sampling of measurement data. CaNNCMs are an important step toward clinical use of data-driven elasticity imaging using AutoP.

Original languageEnglish (US)
Article number8522049
Pages (from-to)1150-1160
Number of pages11
JournalIEEE Transactions on Medical Imaging
Volume38
Issue number5
DOIs
StatePublished - May 2019

Fingerprint

Neural Networks (Computer)
Elasticity
Constitutive models
Neural networks
Imaging techniques
Elasticity Imaging Techniques
Elastic Modulus
Displacement measurement
Force measurement
Noise
Theoretical Models
Materials properties
Elastic moduli
Sampling
Mechanical properties

Keywords

  • Machine learning
  • elastography
  • finite element analysis
  • inverse problems

ASJC Scopus subject areas

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

Cite this

Data-driven elasticity imaging using cartesian neural network constitutive models and the autoprogressive method. / Hoerig, Cameron; Ghaboussi, Jamshid; Insana, Michael.

In: IEEE Transactions on Medical Imaging, Vol. 38, No. 5, 8522049, 05.2019, p. 1150-1160.

Research output: Contribution to journalArticle

@article{b3626af46045434c820f078dd9b82c1b,
title = "Data-driven elasticity imaging using cartesian neural network constitutive models and the autoprogressive method",
abstract = "Quasi-static elasticity imaging techniques rely on model-based mathematical inverse methods to estimate mechanical parameters from force-displacement measurements. These techniques introduce simplifying assumptions that preclude exploration of unknown mechanical properties with potential diagnostic value. We previously reported a data-driven approach to elasticity imaging using artificial neural networks (NNs) that circumvents limitations associated with model-based inverse methods. NN constitutive models can learn stress-strain behavior from force-displacement measurements using the autoprogressive (AutoP) method without prior assumptions of the underlying constitutive model. However, information about internal structure was required. We invented Cartesian NN constitutive models (CaNNCMs) that learn the spatial variations of material properties. We are presenting the first implementation of CaNNCMs trained with AutoP to develop data-driven models of 2-D linear-elastic materials. Both simulated and experimental force-displacement data were used as input to AutoP to show that CaNNCMs are able to model both continuous and discrete material property distributions with no prior information of internal object structure. Furthermore, we demonstrate that CaNNCMs are robust to measurement noise and can reconstruct reasonably accurate Young's modulus images from a sparse sampling of measurement data. CaNNCMs are an important step toward clinical use of data-driven elasticity imaging using AutoP.",
keywords = "Machine learning, elastography, finite element analysis, inverse problems",
author = "Cameron Hoerig and Jamshid Ghaboussi and Michael Insana",
year = "2019",
month = "5",
doi = "10.1109/TMI.2018.2879495",
language = "English (US)",
volume = "38",
pages = "1150--1160",
journal = "IEEE Transactions on Medical Imaging",
issn = "0278-0062",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "5",

}

TY - JOUR

T1 - Data-driven elasticity imaging using cartesian neural network constitutive models and the autoprogressive method

AU - Hoerig, Cameron

AU - Ghaboussi, Jamshid

AU - Insana, Michael

PY - 2019/5

Y1 - 2019/5

N2 - Quasi-static elasticity imaging techniques rely on model-based mathematical inverse methods to estimate mechanical parameters from force-displacement measurements. These techniques introduce simplifying assumptions that preclude exploration of unknown mechanical properties with potential diagnostic value. We previously reported a data-driven approach to elasticity imaging using artificial neural networks (NNs) that circumvents limitations associated with model-based inverse methods. NN constitutive models can learn stress-strain behavior from force-displacement measurements using the autoprogressive (AutoP) method without prior assumptions of the underlying constitutive model. However, information about internal structure was required. We invented Cartesian NN constitutive models (CaNNCMs) that learn the spatial variations of material properties. We are presenting the first implementation of CaNNCMs trained with AutoP to develop data-driven models of 2-D linear-elastic materials. Both simulated and experimental force-displacement data were used as input to AutoP to show that CaNNCMs are able to model both continuous and discrete material property distributions with no prior information of internal object structure. Furthermore, we demonstrate that CaNNCMs are robust to measurement noise and can reconstruct reasonably accurate Young's modulus images from a sparse sampling of measurement data. CaNNCMs are an important step toward clinical use of data-driven elasticity imaging using AutoP.

AB - Quasi-static elasticity imaging techniques rely on model-based mathematical inverse methods to estimate mechanical parameters from force-displacement measurements. These techniques introduce simplifying assumptions that preclude exploration of unknown mechanical properties with potential diagnostic value. We previously reported a data-driven approach to elasticity imaging using artificial neural networks (NNs) that circumvents limitations associated with model-based inverse methods. NN constitutive models can learn stress-strain behavior from force-displacement measurements using the autoprogressive (AutoP) method without prior assumptions of the underlying constitutive model. However, information about internal structure was required. We invented Cartesian NN constitutive models (CaNNCMs) that learn the spatial variations of material properties. We are presenting the first implementation of CaNNCMs trained with AutoP to develop data-driven models of 2-D linear-elastic materials. Both simulated and experimental force-displacement data were used as input to AutoP to show that CaNNCMs are able to model both continuous and discrete material property distributions with no prior information of internal object structure. Furthermore, we demonstrate that CaNNCMs are robust to measurement noise and can reconstruct reasonably accurate Young's modulus images from a sparse sampling of measurement data. CaNNCMs are an important step toward clinical use of data-driven elasticity imaging using AutoP.

KW - Machine learning

KW - elastography

KW - finite element analysis

KW - inverse problems

UR - http://www.scopus.com/inward/record.url?scp=85056158897&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85056158897&partnerID=8YFLogxK

U2 - 10.1109/TMI.2018.2879495

DO - 10.1109/TMI.2018.2879495

M3 - Article

C2 - 30403625

AN - SCOPUS:85056158897

VL - 38

SP - 1150

EP - 1160

JO - IEEE Transactions on Medical Imaging

JF - IEEE Transactions on Medical Imaging

SN - 0278-0062

IS - 5

M1 - 8522049

ER -