Fully automatic segmentation of paraspinal muscles from 3D torso CT images via multi-scale iterative random forest classifications

Naoki Kamiya, Jing Li, Masanori Kume, Hiroshi Fujita, Dinggang Shen, Guoyan Zheng

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

Purpose: To develop and validate a fully automatic method for segmentation of paraspinal muscles from 3D torso CT images. Methods: We propose a novel learning-based method to address this challenging problem. Multi-scale iterative random forest classifications with multi-source information are employed in this study to speed up the segmentation and to improve the accuracy. Here, multi-source images include the original torso CT images and later also the iteratively estimated and refined probability maps of the paraspinal muscles. We validated our method on 20 torso CT data with associated manual segmentation. We randomly partitioned the 20 CT data into two evenly distributed groups and took one group as the training data and the other group as the test data. Results: The proposed method achieved a mean Dice coefficient of 93.0%. It took on average 46.5 s to segment a 3D torso CT image with the size ranging from 512 × 512 × 802 voxels to 512 × 512 × 1031 voxels. Conclusions: Our fully automatic, learning-based method can accurately segment paraspinal muscles from 3D torso CT images. It generates segmentation results that are better than those achieved by the state-of-the-art methods.

Original languageEnglish
JournalInternational journal of computer assisted radiology and surgery
DOIs
Publication statusAccepted/In press - 2018 Jan 1

Fingerprint

Paraspinal Muscles
Torso
Muscle
Learning

Keywords

  • CT
  • Paraspinal muscles
  • Random forest
  • Segmentation

ASJC Scopus subject areas

  • Surgery
  • Biomedical Engineering
  • Radiology Nuclear Medicine and imaging
  • Computer Vision and Pattern Recognition
  • Health Informatics
  • Computer Science Applications
  • Computer Graphics and Computer-Aided Design

Cite this

Fully automatic segmentation of paraspinal muscles from 3D torso CT images via multi-scale iterative random forest classifications. / Kamiya, Naoki; Li, Jing; Kume, Masanori; Fujita, Hiroshi; Shen, Dinggang; Zheng, Guoyan.

In: International journal of computer assisted radiology and surgery, 01.01.2018.

Research output: Contribution to journalArticle

@article{3c68749d09904a688a4475635ac20b9f,
title = "Fully automatic segmentation of paraspinal muscles from 3D torso CT images via multi-scale iterative random forest classifications",
abstract = "Purpose: To develop and validate a fully automatic method for segmentation of paraspinal muscles from 3D torso CT images. Methods: We propose a novel learning-based method to address this challenging problem. Multi-scale iterative random forest classifications with multi-source information are employed in this study to speed up the segmentation and to improve the accuracy. Here, multi-source images include the original torso CT images and later also the iteratively estimated and refined probability maps of the paraspinal muscles. We validated our method on 20 torso CT data with associated manual segmentation. We randomly partitioned the 20 CT data into two evenly distributed groups and took one group as the training data and the other group as the test data. Results: The proposed method achieved a mean Dice coefficient of 93.0{\%}. It took on average 46.5 s to segment a 3D torso CT image with the size ranging from 512 × 512 × 802 voxels to 512 × 512 × 1031 voxels. Conclusions: Our fully automatic, learning-based method can accurately segment paraspinal muscles from 3D torso CT images. It generates segmentation results that are better than those achieved by the state-of-the-art methods.",
keywords = "CT, Paraspinal muscles, Random forest, Segmentation",
author = "Naoki Kamiya and Jing Li and Masanori Kume and Hiroshi Fujita and Dinggang Shen and Guoyan Zheng",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/s11548-018-1852-1",
language = "English",
journal = "International journal of computer assisted radiology and surgery",
issn = "1861-6410",
publisher = "Springer Verlag",

}

TY - JOUR

T1 - Fully automatic segmentation of paraspinal muscles from 3D torso CT images via multi-scale iterative random forest classifications

AU - Kamiya, Naoki

AU - Li, Jing

AU - Kume, Masanori

AU - Fujita, Hiroshi

AU - Shen, Dinggang

AU - Zheng, Guoyan

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Purpose: To develop and validate a fully automatic method for segmentation of paraspinal muscles from 3D torso CT images. Methods: We propose a novel learning-based method to address this challenging problem. Multi-scale iterative random forest classifications with multi-source information are employed in this study to speed up the segmentation and to improve the accuracy. Here, multi-source images include the original torso CT images and later also the iteratively estimated and refined probability maps of the paraspinal muscles. We validated our method on 20 torso CT data with associated manual segmentation. We randomly partitioned the 20 CT data into two evenly distributed groups and took one group as the training data and the other group as the test data. Results: The proposed method achieved a mean Dice coefficient of 93.0%. It took on average 46.5 s to segment a 3D torso CT image with the size ranging from 512 × 512 × 802 voxels to 512 × 512 × 1031 voxels. Conclusions: Our fully automatic, learning-based method can accurately segment paraspinal muscles from 3D torso CT images. It generates segmentation results that are better than those achieved by the state-of-the-art methods.

AB - Purpose: To develop and validate a fully automatic method for segmentation of paraspinal muscles from 3D torso CT images. Methods: We propose a novel learning-based method to address this challenging problem. Multi-scale iterative random forest classifications with multi-source information are employed in this study to speed up the segmentation and to improve the accuracy. Here, multi-source images include the original torso CT images and later also the iteratively estimated and refined probability maps of the paraspinal muscles. We validated our method on 20 torso CT data with associated manual segmentation. We randomly partitioned the 20 CT data into two evenly distributed groups and took one group as the training data and the other group as the test data. Results: The proposed method achieved a mean Dice coefficient of 93.0%. It took on average 46.5 s to segment a 3D torso CT image with the size ranging from 512 × 512 × 802 voxels to 512 × 512 × 1031 voxels. Conclusions: Our fully automatic, learning-based method can accurately segment paraspinal muscles from 3D torso CT images. It generates segmentation results that are better than those achieved by the state-of-the-art methods.

KW - CT

KW - Paraspinal muscles

KW - Random forest

KW - Segmentation

UR - http://www.scopus.com/inward/record.url?scp=85053263313&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85053263313&partnerID=8YFLogxK

U2 - 10.1007/s11548-018-1852-1

DO - 10.1007/s11548-018-1852-1

M3 - Article

C2 - 30173335

AN - SCOPUS:85053263313

JO - International journal of computer assisted radiology and surgery

JF - International journal of computer assisted radiology and surgery

SN - 1861-6410

ER -