Image Registration and Predictive Modeling

Learning the Metric on the Space of Diffeomorphisms

Ayagoz Mussabayeva*, Alexey Kroshnin, Anvar Kurmukov, Yulia Denisova, Li Shen, Shan Cong, Lei Wang, Boris A. Gutman

*Corresponding author for this work

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

Abstract

We present a method for metric optimization in the Large Deformation Diffeomorphic Metric Mapping (LDDMM) framework, by treating the induced Riemannian metric on the space of diffeomorphisms as a kernel in a machine learning context. For simplicity, we choose the kernel Fischer Linear Discriminant Analysis (KLDA) as the framework. Optimizing the kernel parameters in an Expectation-Maximization framework, we define model fidelity via the hinge loss of the decision function. The resulting algorithm optimizes the parameters of the LDDMM norm-inducing differential operator as a solution to a group-wise registration and classification problem. In practice, this may lead to a biology-aware registration, focusing its attention on the predictive task at hand such as identifying the effects of disease. We first tested our algorithm on a synthetic dataset, showing that our parameter selection improves registration quality and classification accuracy. We then tested the algorithm on 3D subcortical shapes from the Schizophrenia cohort Schizconnect. Our Schizophrenia-Control predictive model showed significant improvement in ROC AUC compared to baseline parameters.

Original languageEnglish (US)
Title of host publicationShape in Medical Imaging - International Workshop, ShapeMI 2018, Held in Conjunction with MICCAI 2018, Proceedings
EditorsHervé Lombaert, Beatriz Paniagua, Bernhard Egger, Marcel Lüthi, Martin Reuter, Christian Wachinger
PublisherSpringer Verlag
Pages160-168
Number of pages9
ISBN (Print)9783030047467
DOIs
StatePublished - Jan 1 2018
EventInternational Workshop on Shape in Medical Imaging, ShapeMI 2018 held in conjunction with 21st International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2018 - Granada, Spain
Duration: Sep 20 2018Sep 20 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11167 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

OtherInternational Workshop on Shape in Medical Imaging, ShapeMI 2018 held in conjunction with 21st International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2018
CountrySpain
CityGranada
Period9/20/189/20/18

Fingerprint

Predictive Modeling
Image registration
Image Registration
Diffeomorphisms
Registration
Large Deformation
kernel
Metric
Expectation Maximization
3D shape
Parameter Selection
Model predictive control
Model Predictive Control
Discriminant analysis
Hinges
Riemannian Metric
Discriminant Analysis
Classification Problems
Fidelity
Biology

Keywords

  • Expectation Maximization
  • Image registration
  • LDDMM
  • Machine learning
  • Metric learning
  • Subcortical shape

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Mussabayeva, A., Kroshnin, A., Kurmukov, A., Denisova, Y., Shen, L., Cong, S., ... Gutman, B. A. (2018). Image Registration and Predictive Modeling: Learning the Metric on the Space of Diffeomorphisms. In H. Lombaert, B. Paniagua, B. Egger, M. Lüthi, M. Reuter, & C. Wachinger (Eds.), Shape in Medical Imaging - International Workshop, ShapeMI 2018, Held in Conjunction with MICCAI 2018, Proceedings (pp. 160-168). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11167 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-04747-4_15
Mussabayeva, Ayagoz ; Kroshnin, Alexey ; Kurmukov, Anvar ; Denisova, Yulia ; Shen, Li ; Cong, Shan ; Wang, Lei ; Gutman, Boris A. / Image Registration and Predictive Modeling : Learning the Metric on the Space of Diffeomorphisms. Shape in Medical Imaging - International Workshop, ShapeMI 2018, Held in Conjunction with MICCAI 2018, Proceedings. editor / Hervé Lombaert ; Beatriz Paniagua ; Bernhard Egger ; Marcel Lüthi ; Martin Reuter ; Christian Wachinger. Springer Verlag, 2018. pp. 160-168 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{ffb89d990fdf43ba81cd0dbc0f0b23e6,
title = "Image Registration and Predictive Modeling: Learning the Metric on the Space of Diffeomorphisms",
abstract = "We present a method for metric optimization in the Large Deformation Diffeomorphic Metric Mapping (LDDMM) framework, by treating the induced Riemannian metric on the space of diffeomorphisms as a kernel in a machine learning context. For simplicity, we choose the kernel Fischer Linear Discriminant Analysis (KLDA) as the framework. Optimizing the kernel parameters in an Expectation-Maximization framework, we define model fidelity via the hinge loss of the decision function. The resulting algorithm optimizes the parameters of the LDDMM norm-inducing differential operator as a solution to a group-wise registration and classification problem. In practice, this may lead to a biology-aware registration, focusing its attention on the predictive task at hand such as identifying the effects of disease. We first tested our algorithm on a synthetic dataset, showing that our parameter selection improves registration quality and classification accuracy. We then tested the algorithm on 3D subcortical shapes from the Schizophrenia cohort Schizconnect. Our Schizophrenia-Control predictive model showed significant improvement in ROC AUC compared to baseline parameters.",
keywords = "Expectation Maximization, Image registration, LDDMM, Machine learning, Metric learning, Subcortical shape",
author = "Ayagoz Mussabayeva and Alexey Kroshnin and Anvar Kurmukov and Yulia Denisova and Li Shen and Shan Cong and Lei Wang and Gutman, {Boris A.}",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-3-030-04747-4_15",
language = "English (US)",
isbn = "9783030047467",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "160--168",
editor = "Herv{\'e} Lombaert and Beatriz Paniagua and Bernhard Egger and Marcel L{\"u}thi and Martin Reuter and Christian Wachinger",
booktitle = "Shape in Medical Imaging - International Workshop, ShapeMI 2018, Held in Conjunction with MICCAI 2018, Proceedings",
address = "Germany",

}

Mussabayeva, A, Kroshnin, A, Kurmukov, A, Denisova, Y, Shen, L, Cong, S, Wang, L & Gutman, BA 2018, Image Registration and Predictive Modeling: Learning the Metric on the Space of Diffeomorphisms. in H Lombaert, B Paniagua, B Egger, M Lüthi, M Reuter & C Wachinger (eds), Shape in Medical Imaging - International Workshop, ShapeMI 2018, Held in Conjunction with MICCAI 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11167 LNCS, Springer Verlag, pp. 160-168, International Workshop on Shape in Medical Imaging, ShapeMI 2018 held in conjunction with 21st International Conference on Medical Image Computing and Computer Assisted Intervention, MICCAI 2018, Granada, Spain, 9/20/18. https://doi.org/10.1007/978-3-030-04747-4_15

Image Registration and Predictive Modeling : Learning the Metric on the Space of Diffeomorphisms. / Mussabayeva, Ayagoz; Kroshnin, Alexey; Kurmukov, Anvar; Denisova, Yulia; Shen, Li; Cong, Shan; Wang, Lei; Gutman, Boris A.

Shape in Medical Imaging - International Workshop, ShapeMI 2018, Held in Conjunction with MICCAI 2018, Proceedings. ed. / Hervé Lombaert; Beatriz Paniagua; Bernhard Egger; Marcel Lüthi; Martin Reuter; Christian Wachinger. Springer Verlag, 2018. p. 160-168 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11167 LNCS).

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

TY - GEN

T1 - Image Registration and Predictive Modeling

T2 - Learning the Metric on the Space of Diffeomorphisms

AU - Mussabayeva, Ayagoz

AU - Kroshnin, Alexey

AU - Kurmukov, Anvar

AU - Denisova, Yulia

AU - Shen, Li

AU - Cong, Shan

AU - Wang, Lei

AU - Gutman, Boris A.

PY - 2018/1/1

Y1 - 2018/1/1

N2 - We present a method for metric optimization in the Large Deformation Diffeomorphic Metric Mapping (LDDMM) framework, by treating the induced Riemannian metric on the space of diffeomorphisms as a kernel in a machine learning context. For simplicity, we choose the kernel Fischer Linear Discriminant Analysis (KLDA) as the framework. Optimizing the kernel parameters in an Expectation-Maximization framework, we define model fidelity via the hinge loss of the decision function. The resulting algorithm optimizes the parameters of the LDDMM norm-inducing differential operator as a solution to a group-wise registration and classification problem. In practice, this may lead to a biology-aware registration, focusing its attention on the predictive task at hand such as identifying the effects of disease. We first tested our algorithm on a synthetic dataset, showing that our parameter selection improves registration quality and classification accuracy. We then tested the algorithm on 3D subcortical shapes from the Schizophrenia cohort Schizconnect. Our Schizophrenia-Control predictive model showed significant improvement in ROC AUC compared to baseline parameters.

AB - We present a method for metric optimization in the Large Deformation Diffeomorphic Metric Mapping (LDDMM) framework, by treating the induced Riemannian metric on the space of diffeomorphisms as a kernel in a machine learning context. For simplicity, we choose the kernel Fischer Linear Discriminant Analysis (KLDA) as the framework. Optimizing the kernel parameters in an Expectation-Maximization framework, we define model fidelity via the hinge loss of the decision function. The resulting algorithm optimizes the parameters of the LDDMM norm-inducing differential operator as a solution to a group-wise registration and classification problem. In practice, this may lead to a biology-aware registration, focusing its attention on the predictive task at hand such as identifying the effects of disease. We first tested our algorithm on a synthetic dataset, showing that our parameter selection improves registration quality and classification accuracy. We then tested the algorithm on 3D subcortical shapes from the Schizophrenia cohort Schizconnect. Our Schizophrenia-Control predictive model showed significant improvement in ROC AUC compared to baseline parameters.

KW - Expectation Maximization

KW - Image registration

KW - LDDMM

KW - Machine learning

KW - Metric learning

KW - Subcortical shape

UR - http://www.scopus.com/inward/record.url?scp=85057381599&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85057381599&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-04747-4_15

DO - 10.1007/978-3-030-04747-4_15

M3 - Conference contribution

SN - 9783030047467

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 160

EP - 168

BT - Shape in Medical Imaging - International Workshop, ShapeMI 2018, Held in Conjunction with MICCAI 2018, Proceedings

A2 - Lombaert, Hervé

A2 - Paniagua, Beatriz

A2 - Egger, Bernhard

A2 - Lüthi, Marcel

A2 - Reuter, Martin

A2 - Wachinger, Christian

PB - Springer Verlag

ER -

Mussabayeva A, Kroshnin A, Kurmukov A, Denisova Y, Shen L, Cong S et al. Image Registration and Predictive Modeling: Learning the Metric on the Space of Diffeomorphisms. In Lombaert H, Paniagua B, Egger B, Lüthi M, Reuter M, Wachinger C, editors, Shape in Medical Imaging - International Workshop, ShapeMI 2018, Held in Conjunction with MICCAI 2018, Proceedings. Springer Verlag. 2018. p. 160-168. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-04747-4_15