Comparison of automatic vessel segmentation techniques for whole body magnetic resonance angiography with limited ground truth data

Andrew McNeil (Lead / Corresponding author), Giulio Degano, Ian Poole, Graeme Houston, Emanuele Trucco

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

1 Citation (Scopus)

Abstract

This work is part of a project aimed at automatically detecting vascular disease in whole body magnetic resonance angiograms (WBMRA). Here we present a comparison of four techniques for automatic artery segmentation in WBMRA data volumes; active contours, two “vesselness” filter approaches (the Frangi filter and Optimally Oriented Flux (OOF)) and a convolutional neural network (Convnet) trained for voxel-wise classification. Their performance was assessed on three manually segmented WBMRA datasets, comparing the maximum Dice Similarity Coefficient (DSC) achieved by each method. Our results show that, in the presence of limited training data, OOF performs best for our three patients, achieving a mean DSC of 0.71 across all patients. By comparison, the 3D Convnet achieved a mean DSC of 0.63. We discuss the potential reasons for these differences, and the implications it has for the automated segmentation of arteries in large WBMRA datasets, where ground truth data is often limited and there are currently no pre-trained 3D Convnet models available, requiring models to be trained from scratch. To the best of our knowledge this is the first comparison of these automated vessel segmentation techniques for WBMRA data, and the first quantitative results of applying a Convnet to vessel segmentation in WBMRA, for which no public sets of manually annotated vascular networks currently exist.

Original languageEnglish
Title of host publicationMedical Image Understanding and Analysis
Subtitle of host publication21st Annual Conference, MIUA 2017, Edinburgh, UK, July 11–13, 2017, Proceedings
EditorsMaria Valdez Hernandez, Victor Gonzalez-Castro
Place of PublicationSwitzerland
PublisherSpringer
Pages144-155
Number of pages12
ISBN (Electronic)9783319609645
ISBN (Print)9783319609638
DOIs
Publication statusPublished - 2017
Event21st Annual Conference on Medical Image Understanding and Analysis, MIUA 2017 - Edinburgh, United Kingdom
Duration: 11 Jul 201713 Jul 2017

Publication series

NameCommunications in Computer and Information Science
PublisherSpringer
Volume723
ISSN (Print)1865-0929
ISSN (Electronic)1865-0937

Conference

Conference21st Annual Conference on Medical Image Understanding and Analysis, MIUA 2017
CountryUnited Kingdom
CityEdinburgh
Period11/07/1713/07/17

Fingerprint

Angiography
Magnetic resonance
Neural networks
Fluxes

Cite this

McNeil, A., Degano, G., Poole, I., Houston, G., & Trucco, E. (2017). Comparison of automatic vessel segmentation techniques for whole body magnetic resonance angiography with limited ground truth data. In M. V. Hernandez, & V. Gonzalez-Castro (Eds.), Medical Image Understanding and Analysis: 21st Annual Conference, MIUA 2017, Edinburgh, UK, July 11–13, 2017, Proceedings (pp. 144-155). (Communications in Computer and Information Science; Vol. 723). Switzerland: Springer . https://doi.org/10.1007/978-3-319-60964-5_13
McNeil, Andrew ; Degano, Giulio ; Poole, Ian ; Houston, Graeme ; Trucco, Emanuele. / Comparison of automatic vessel segmentation techniques for whole body magnetic resonance angiography with limited ground truth data. Medical Image Understanding and Analysis: 21st Annual Conference, MIUA 2017, Edinburgh, UK, July 11–13, 2017, Proceedings. editor / Maria Valdez Hernandez ; Victor Gonzalez-Castro. Switzerland : Springer , 2017. pp. 144-155 (Communications in Computer and Information Science).
@inproceedings{a384c5f24bd0410e8645935be94a4a15,
title = "Comparison of automatic vessel segmentation techniques for whole body magnetic resonance angiography with limited ground truth data",
abstract = "This work is part of a project aimed at automatically detecting vascular disease in whole body magnetic resonance angiograms (WBMRA). Here we present a comparison of four techniques for automatic artery segmentation in WBMRA data volumes; active contours, two “vesselness” filter approaches (the Frangi filter and Optimally Oriented Flux (OOF)) and a convolutional neural network (Convnet) trained for voxel-wise classification. Their performance was assessed on three manually segmented WBMRA datasets, comparing the maximum Dice Similarity Coefficient (DSC) achieved by each method. Our results show that, in the presence of limited training data, OOF performs best for our three patients, achieving a mean DSC of 0.71 across all patients. By comparison, the 3D Convnet achieved a mean DSC of 0.63. We discuss the potential reasons for these differences, and the implications it has for the automated segmentation of arteries in large WBMRA datasets, where ground truth data is often limited and there are currently no pre-trained 3D Convnet models available, requiring models to be trained from scratch. To the best of our knowledge this is the first comparison of these automated vessel segmentation techniques for WBMRA data, and the first quantitative results of applying a Convnet to vessel segmentation in WBMRA, for which no public sets of manually annotated vascular networks currently exist.",
author = "Andrew McNeil and Giulio Degano and Ian Poole and Graeme Houston and Emanuele Trucco",
note = "{\circledC} Springer International Publishing AG 2017",
year = "2017",
doi = "10.1007/978-3-319-60964-5_13",
language = "English",
isbn = "9783319609638",
series = "Communications in Computer and Information Science",
publisher = "Springer",
pages = "144--155",
editor = "Hernandez, {Maria Valdez} and Victor Gonzalez-Castro",
booktitle = "Medical Image Understanding and Analysis",

}

McNeil, A, Degano, G, Poole, I, Houston, G & Trucco, E 2017, Comparison of automatic vessel segmentation techniques for whole body magnetic resonance angiography with limited ground truth data. in MV Hernandez & V Gonzalez-Castro (eds), Medical Image Understanding and Analysis: 21st Annual Conference, MIUA 2017, Edinburgh, UK, July 11–13, 2017, Proceedings. Communications in Computer and Information Science, vol. 723, Springer , Switzerland, pp. 144-155, 21st Annual Conference on Medical Image Understanding and Analysis, MIUA 2017, Edinburgh, United Kingdom, 11/07/17. https://doi.org/10.1007/978-3-319-60964-5_13

Comparison of automatic vessel segmentation techniques for whole body magnetic resonance angiography with limited ground truth data. / McNeil, Andrew (Lead / Corresponding author); Degano, Giulio; Poole, Ian; Houston, Graeme; Trucco, Emanuele.

Medical Image Understanding and Analysis: 21st Annual Conference, MIUA 2017, Edinburgh, UK, July 11–13, 2017, Proceedings. ed. / Maria Valdez Hernandez; Victor Gonzalez-Castro. Switzerland : Springer , 2017. p. 144-155 (Communications in Computer and Information Science; Vol. 723).

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

TY - GEN

T1 - Comparison of automatic vessel segmentation techniques for whole body magnetic resonance angiography with limited ground truth data

AU - McNeil, Andrew

AU - Degano, Giulio

AU - Poole, Ian

AU - Houston, Graeme

AU - Trucco, Emanuele

N1 - © Springer International Publishing AG 2017

PY - 2017

Y1 - 2017

N2 - This work is part of a project aimed at automatically detecting vascular disease in whole body magnetic resonance angiograms (WBMRA). Here we present a comparison of four techniques for automatic artery segmentation in WBMRA data volumes; active contours, two “vesselness” filter approaches (the Frangi filter and Optimally Oriented Flux (OOF)) and a convolutional neural network (Convnet) trained for voxel-wise classification. Their performance was assessed on three manually segmented WBMRA datasets, comparing the maximum Dice Similarity Coefficient (DSC) achieved by each method. Our results show that, in the presence of limited training data, OOF performs best for our three patients, achieving a mean DSC of 0.71 across all patients. By comparison, the 3D Convnet achieved a mean DSC of 0.63. We discuss the potential reasons for these differences, and the implications it has for the automated segmentation of arteries in large WBMRA datasets, where ground truth data is often limited and there are currently no pre-trained 3D Convnet models available, requiring models to be trained from scratch. To the best of our knowledge this is the first comparison of these automated vessel segmentation techniques for WBMRA data, and the first quantitative results of applying a Convnet to vessel segmentation in WBMRA, for which no public sets of manually annotated vascular networks currently exist.

AB - This work is part of a project aimed at automatically detecting vascular disease in whole body magnetic resonance angiograms (WBMRA). Here we present a comparison of four techniques for automatic artery segmentation in WBMRA data volumes; active contours, two “vesselness” filter approaches (the Frangi filter and Optimally Oriented Flux (OOF)) and a convolutional neural network (Convnet) trained for voxel-wise classification. Their performance was assessed on three manually segmented WBMRA datasets, comparing the maximum Dice Similarity Coefficient (DSC) achieved by each method. Our results show that, in the presence of limited training data, OOF performs best for our three patients, achieving a mean DSC of 0.71 across all patients. By comparison, the 3D Convnet achieved a mean DSC of 0.63. We discuss the potential reasons for these differences, and the implications it has for the automated segmentation of arteries in large WBMRA datasets, where ground truth data is often limited and there are currently no pre-trained 3D Convnet models available, requiring models to be trained from scratch. To the best of our knowledge this is the first comparison of these automated vessel segmentation techniques for WBMRA data, and the first quantitative results of applying a Convnet to vessel segmentation in WBMRA, for which no public sets of manually annotated vascular networks currently exist.

UR - http://www.scopus.com/inward/record.url?scp=85022209981&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-60964-5_13

DO - 10.1007/978-3-319-60964-5_13

M3 - Conference contribution

SN - 9783319609638

T3 - Communications in Computer and Information Science

SP - 144

EP - 155

BT - Medical Image Understanding and Analysis

A2 - Hernandez, Maria Valdez

A2 - Gonzalez-Castro, Victor

PB - Springer

CY - Switzerland

ER -

McNeil A, Degano G, Poole I, Houston G, Trucco E. Comparison of automatic vessel segmentation techniques for whole body magnetic resonance angiography with limited ground truth data. In Hernandez MV, Gonzalez-Castro V, editors, Medical Image Understanding and Analysis: 21st Annual Conference, MIUA 2017, Edinburgh, UK, July 11–13, 2017, Proceedings. Switzerland: Springer . 2017. p. 144-155. (Communications in Computer and Information Science). https://doi.org/10.1007/978-3-319-60964-5_13