A geodesic landmark shooting algorithm for template matching and its applications

Roberto Camassa, Dongyang Kuang, Long Lee

Research output: Research - peer-reviewArticle

Abstract

We present an efficient landmark shooting algorithm for template matching and its applications. The novelties of the algorithm include the use of a constant matrix to update the search direction of the geodesic shooting, instead of the traditional methods of forward-backward integration for updating the gradient or Newton’s optimization, and the use of a nonsmooth conic kernel for the particle system that accelerates the convergence of matching. To investigate the usage of the output quantities computed along the warping algorithm, such as the Hamiltonian metric and the momentum field, we introduce a multiscale decomposition method that separates the scales/components of the momentum and the Hamiltonian metric associated with the deformation. We numerically explore the potential of using the decomposed Hamiltonian metric and momentum vectors as input feature vectors into neural networks for clustering/classification analysis. The results of our numerical experiments are encouraging.

LanguageEnglish (US)
Pages303-334
Number of pages32
JournalSIAM Journal on Imaging Sciences
Volume10
Issue number1
DOIs
StatePublished - 2017

Fingerprint

Template Matching
Shooting
Landmarks
Geodesic
Momentum
Metric
Hamiltonians
Template matching
Multiscale Methods
Warping
Particle System
Decomposition Method
Feature Vector
Accelerate
Updating
Update
Numerical Experiment
Clustering
Neural Networks
kernel

Keywords

  • Deformation
  • Geodesic shooting
  • Hamiltonian metric
  • Landmark
  • Multiscale decomposition
  • Neural networks
  • Particle system
  • Template matching

ASJC Scopus subject areas

  • Mathematics(all)
  • Applied Mathematics

Cite this

A geodesic landmark shooting algorithm for template matching and its applications. / Camassa, Roberto; Kuang, Dongyang; Lee, Long.

In: SIAM Journal on Imaging Sciences, Vol. 10, No. 1, 2017, p. 303-334.

Research output: Research - peer-reviewArticle

@article{84168cf69eb64de98e29f34593564b3a,
title = "A geodesic landmark shooting algorithm for template matching and its applications",
abstract = "We present an efficient landmark shooting algorithm for template matching and its applications. The novelties of the algorithm include the use of a constant matrix to update the search direction of the geodesic shooting, instead of the traditional methods of forward-backward integration for updating the gradient or Newton’s optimization, and the use of a nonsmooth conic kernel for the particle system that accelerates the convergence of matching. To investigate the usage of the output quantities computed along the warping algorithm, such as the Hamiltonian metric and the momentum field, we introduce a multiscale decomposition method that separates the scales/components of the momentum and the Hamiltonian metric associated with the deformation. We numerically explore the potential of using the decomposed Hamiltonian metric and momentum vectors as input feature vectors into neural networks for clustering/classification analysis. The results of our numerical experiments are encouraging.",
keywords = "Deformation, Geodesic shooting, Hamiltonian metric, Landmark, Multiscale decomposition, Neural networks, Particle system, Template matching",
author = "Roberto Camassa and Dongyang Kuang and Long Lee",
year = "2017",
doi = "10.1137/15M104373X",
volume = "10",
pages = "303--334",
journal = "SIAM Journal on Imaging Sciences",
issn = "1936-4954",
publisher = "Society for Industrial and Applied Mathematics Publications",
number = "1",

}

TY - JOUR

T1 - A geodesic landmark shooting algorithm for template matching and its applications

AU - Camassa,Roberto

AU - Kuang,Dongyang

AU - Lee,Long

PY - 2017

Y1 - 2017

N2 - We present an efficient landmark shooting algorithm for template matching and its applications. The novelties of the algorithm include the use of a constant matrix to update the search direction of the geodesic shooting, instead of the traditional methods of forward-backward integration for updating the gradient or Newton’s optimization, and the use of a nonsmooth conic kernel for the particle system that accelerates the convergence of matching. To investigate the usage of the output quantities computed along the warping algorithm, such as the Hamiltonian metric and the momentum field, we introduce a multiscale decomposition method that separates the scales/components of the momentum and the Hamiltonian metric associated with the deformation. We numerically explore the potential of using the decomposed Hamiltonian metric and momentum vectors as input feature vectors into neural networks for clustering/classification analysis. The results of our numerical experiments are encouraging.

AB - We present an efficient landmark shooting algorithm for template matching and its applications. The novelties of the algorithm include the use of a constant matrix to update the search direction of the geodesic shooting, instead of the traditional methods of forward-backward integration for updating the gradient or Newton’s optimization, and the use of a nonsmooth conic kernel for the particle system that accelerates the convergence of matching. To investigate the usage of the output quantities computed along the warping algorithm, such as the Hamiltonian metric and the momentum field, we introduce a multiscale decomposition method that separates the scales/components of the momentum and the Hamiltonian metric associated with the deformation. We numerically explore the potential of using the decomposed Hamiltonian metric and momentum vectors as input feature vectors into neural networks for clustering/classification analysis. The results of our numerical experiments are encouraging.

KW - Deformation

KW - Geodesic shooting

KW - Hamiltonian metric

KW - Landmark

KW - Multiscale decomposition

KW - Neural networks

KW - Particle system

KW - Template matching

UR - http://www.scopus.com/inward/record.url?scp=85016580255&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85016580255&partnerID=8YFLogxK

U2 - 10.1137/15M104373X

DO - 10.1137/15M104373X

M3 - Article

VL - 10

SP - 303

EP - 334

JO - SIAM Journal on Imaging Sciences

T2 - SIAM Journal on Imaging Sciences

JF - SIAM Journal on Imaging Sciences

SN - 1936-4954

IS - 1

ER -