3D segmentation for multi-organs in CT images

Mariusz Bajger; Gobert Lee; Martin Caon

doi:10.5565/rev/elcvia.516

3D segmentation for multi-organs in CT images

Mariusz Bajger, Gobert Lee, Martin Caon

Research output: Contribution to journal › Article › peer-review

15 Citations (Scopus)

43 Downloads (Pure)

Abstract

The study addresses the challenging problemof automatic segmentation of the human anatomy needed for radiation dose calculations. Three-dimensional extensions of two well-known stateof-the art segmentation techniques are proposed and tested for usefulness on a set of clinical CT images. The new techniques are 3D Statistical Region Merging (3D-SRM) and 3D Efficient Graph-based Segmentation (3D-EGS). Segmentations of eight representative tissues (lungs, stomach, liver, heart, kidneys, spleen, bones and the spinal cord) were tested for accuracy using the Dice index, the Hausdorff distance and the H_t index. The 3D-SRM outperformed 3D-EGS producing the average (across the 8 tissues) Dice index, the Hausdorff distance, and the H₂ of 0.89, 12.5 mm and 0.93, respectively.

Original language	English
Pages (from-to)	13-27
Number of pages	15
Journal	Electronic Letters on Computer Vision and Image Analysis
Volume	12
Issue number	2
DOIs	https://doi.org/10.5565/rev/elcvia.516
Publication status	Published - 2013

Keywords

Efficient graph-based segmentation
Full-body CT
Image segmentation
Statistical region merging
Voxel model

Access to Document

10.5565/rev/elcvia.516Licence: CC BY-NC-ND

Bajger_3D_P2013Final published version, 164 KBLicence: CC BY-NC-ND

Cite this

@article{77fff431473e44a7a07dc4715d2051ed,

title = "3D segmentation for multi-organs in CT images",

abstract = "The study addresses the challenging problemof automatic segmentation of the human anatomy needed for radiation dose calculations. Three-dimensional extensions of two well-known stateof-the art segmentation techniques are proposed and tested for usefulness on a set of clinical CT images. The new techniques are 3D Statistical Region Merging (3D-SRM) and 3D Efficient Graph-based Segmentation (3D-EGS). Segmentations of eight representative tissues (lungs, stomach, liver, heart, kidneys, spleen, bones and the spinal cord) were tested for accuracy using the Dice index, the Hausdorff distance and the Ht index. The 3D-SRM outperformed 3D-EGS producing the average (across the 8 tissues) Dice index, the Hausdorff distance, and the H2 of 0.89, 12.5 mm and 0.93, respectively.",

keywords = "Efficient graph-based segmentation, Full-body CT, Image segmentation, Statistical region merging, Voxel model",

author = "Mariusz Bajger and Gobert Lee and Martin Caon",

year = "2013",

doi = "10.5565/rev/elcvia.516",

language = "English",

volume = "12",

pages = "13--27",

journal = "Electronic Letters on Computer Vision and Image Analysis",

issn = "1577-5097",

publisher = "Centre de Visio per Computador",

number = "2",

}

TY - JOUR

T1 - 3D segmentation for multi-organs in CT images

AU - Bajger, Mariusz

AU - Lee, Gobert

AU - Caon, Martin

PY - 2013

Y1 - 2013

N2 - The study addresses the challenging problemof automatic segmentation of the human anatomy needed for radiation dose calculations. Three-dimensional extensions of two well-known stateof-the art segmentation techniques are proposed and tested for usefulness on a set of clinical CT images. The new techniques are 3D Statistical Region Merging (3D-SRM) and 3D Efficient Graph-based Segmentation (3D-EGS). Segmentations of eight representative tissues (lungs, stomach, liver, heart, kidneys, spleen, bones and the spinal cord) were tested for accuracy using the Dice index, the Hausdorff distance and the Ht index. The 3D-SRM outperformed 3D-EGS producing the average (across the 8 tissues) Dice index, the Hausdorff distance, and the H2 of 0.89, 12.5 mm and 0.93, respectively.

AB - The study addresses the challenging problemof automatic segmentation of the human anatomy needed for radiation dose calculations. Three-dimensional extensions of two well-known stateof-the art segmentation techniques are proposed and tested for usefulness on a set of clinical CT images. The new techniques are 3D Statistical Region Merging (3D-SRM) and 3D Efficient Graph-based Segmentation (3D-EGS). Segmentations of eight representative tissues (lungs, stomach, liver, heart, kidneys, spleen, bones and the spinal cord) were tested for accuracy using the Dice index, the Hausdorff distance and the Ht index. The 3D-SRM outperformed 3D-EGS producing the average (across the 8 tissues) Dice index, the Hausdorff distance, and the H2 of 0.89, 12.5 mm and 0.93, respectively.

KW - Efficient graph-based segmentation

KW - Full-body CT

KW - Image segmentation

KW - Statistical region merging

KW - Voxel model

UR - http://www.scopus.com/inward/record.url?scp=84877314302&partnerID=8YFLogxK

U2 - 10.5565/rev/elcvia.516

DO - 10.5565/rev/elcvia.516

M3 - Article

SN - 1577-5097

VL - 12

SP - 13

EP - 27

JO - Electronic Letters on Computer Vision and Image Analysis

JF - Electronic Letters on Computer Vision and Image Analysis

IS - 2

ER -

3D segmentation for multi-organs in CT images

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this