Increasing Accuracy of Optimal Surfaces Using Min-marginal Energies

Jens Petersen; Andres M Arias-Lorza; Raghavendra Selvan; Daniel Bos; Aad van der Lugt; Jesper H Pedersen; Mads Nielsen; Marleen de Bruijne

doi:10.1109/TMI.2018.2890386

Increasing Accuracy of Optimal Surfaces Using Min-marginal Energies

Jens Petersen, Andres M Arias-Lorza, Raghavendra Selvan, Daniel Bos, Aad van der Lugt, Jesper H Pedersen, Mads Nielsen, Marleen de Bruijne

1 Citation (Scopus)

Abstract

Optimal surface methods are a class of graph cut methods posing surface estimation as an n-ary ordered labeling problem. They are used in medical imaging to find interacting and layered surfaces optimally and in low order polynomial time. Representing continuous surfaces with discrete sets of labels, however, leads to discretization errors and, if graph representations are made dense, excessive memory usage. Limiting memory usage and computation time of graph cut methods are important and graphs that locally adapt to the problem has been proposed as a solution. Min-marginal energies computed using dynamic graph cuts offer a way to estimate solution uncertainty and these uncertainties have been used to decide where graphs should be adapted. Adaptive graphs, however, introduce extra parameters, complexity, and heuristics. We propose a way to use min-marginal energies to estimate continuous solution labels that does not introduce extra parameters and show empirically on synthetic and medical imaging datasets that it leads to improved accuracy. The increase in accuracy was consistent and in many cases comparable with accuracy otherwise obtained with graphs up to eight times denser, but with proportionally less memory usage and improvements in computation time.

Original language	English
Article number	8599009
Journal	IEEE Transactions on Medical Imaging
Volume	38
Issue number	7
Pages (from-to)	1559-1568
ISSN	0278-0062
DOIs	https://doi.org/10.1109/TMI.2018.2890386
Publication status	Published - Jul 2019

Access to Document

10.1109/TMI.2018.2890386

Cite this

@article{af1a420289f7476a8d7478cb7484e2cb,

title = "Increasing Accuracy of Optimal Surfaces Using Min-marginal Energies",

abstract = "Optimal surface methods are a class of graph cut methods posing surface estimation as an n-ary ordered labeling problem. They are used in medical imaging to find interacting and layered surfaces optimally and in low order polynomial time. Representing continuous surfaces with discrete sets of labels, however, leads to discretization errors and, if graph representations are made dense, excessive memory usage. Limiting memory usage and computation time of graph cut methods are important and graphs that locally adapt to the problem has been proposed as a solution. Min-marginal energies computed using dynamic graph cuts offer a way to estimate solution uncertainty and these uncertainties have been used to decide where graphs should be adapted. Adaptive graphs, however, introduce extra parameters, complexity, and heuristics. We propose a way to use min-marginal energies to estimate continuous solution labels that does not introduce extra parameters and show empirically on synthetic and medical imaging datasets that it leads to improved accuracy. The increase in accuracy was consistent and in many cases comparable with accuracy otherwise obtained with graphs up to eight times denser, but with proportionally less memory usage and improvements in computation time.",

author = "Jens Petersen and Arias-Lorza, {Andres M} and Raghavendra Selvan and Daniel Bos and {van der Lugt}, Aad and Pedersen, {Jesper H} and Mads Nielsen and {de Bruijne}, Marleen",

year = "2019",

month = jul,

doi = "10.1109/TMI.2018.2890386",

language = "English",

volume = "38",

pages = "1559--1568",

journal = "IEEE Transactions on Medical Imaging",

issn = "0278-0062",

publisher = "Institute of Electrical and Electronics Engineers",

number = "7",

}

TY - JOUR

T1 - Increasing Accuracy of Optimal Surfaces Using Min-marginal Energies

AU - Petersen, Jens

AU - Arias-Lorza, Andres M

AU - Selvan, Raghavendra

AU - Bos, Daniel

AU - van der Lugt, Aad

AU - Pedersen, Jesper H

AU - Nielsen, Mads

AU - de Bruijne, Marleen

PY - 2019/7

Y1 - 2019/7

N2 - Optimal surface methods are a class of graph cut methods posing surface estimation as an n-ary ordered labeling problem. They are used in medical imaging to find interacting and layered surfaces optimally and in low order polynomial time. Representing continuous surfaces with discrete sets of labels, however, leads to discretization errors and, if graph representations are made dense, excessive memory usage. Limiting memory usage and computation time of graph cut methods are important and graphs that locally adapt to the problem has been proposed as a solution. Min-marginal energies computed using dynamic graph cuts offer a way to estimate solution uncertainty and these uncertainties have been used to decide where graphs should be adapted. Adaptive graphs, however, introduce extra parameters, complexity, and heuristics. We propose a way to use min-marginal energies to estimate continuous solution labels that does not introduce extra parameters and show empirically on synthetic and medical imaging datasets that it leads to improved accuracy. The increase in accuracy was consistent and in many cases comparable with accuracy otherwise obtained with graphs up to eight times denser, but with proportionally less memory usage and improvements in computation time.

AB - Optimal surface methods are a class of graph cut methods posing surface estimation as an n-ary ordered labeling problem. They are used in medical imaging to find interacting and layered surfaces optimally and in low order polynomial time. Representing continuous surfaces with discrete sets of labels, however, leads to discretization errors and, if graph representations are made dense, excessive memory usage. Limiting memory usage and computation time of graph cut methods are important and graphs that locally adapt to the problem has been proposed as a solution. Min-marginal energies computed using dynamic graph cuts offer a way to estimate solution uncertainty and these uncertainties have been used to decide where graphs should be adapted. Adaptive graphs, however, introduce extra parameters, complexity, and heuristics. We propose a way to use min-marginal energies to estimate continuous solution labels that does not introduce extra parameters and show empirically on synthetic and medical imaging datasets that it leads to improved accuracy. The increase in accuracy was consistent and in many cases comparable with accuracy otherwise obtained with graphs up to eight times denser, but with proportionally less memory usage and improvements in computation time.

U2 - 10.1109/TMI.2018.2890386

DO - 10.1109/TMI.2018.2890386

M3 - Journal article

C2 - 30605096

SN - 0278-0062

VL - 38

SP - 1559

EP - 1568

JO - IEEE Transactions on Medical Imaging

JF - IEEE Transactions on Medical Imaging

IS - 7

M1 - 8599009

ER -

Increasing Accuracy of Optimal Surfaces Using Min-marginal Energies

Abstract

Access to Document

Fingerprint

Cite this