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 ShenEmail author
  • Guoyan ZhengEmail author
Original Article



To develop and validate a fully automatic method for segmentation of paraspinal muscles from 3D torso CT images.


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.


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 \(512 \times 512 \times 802\) voxels to \(512 \times 512 \times 1031\) voxels.


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.


Paraspinal muscles CT Segmentation Random forest 



This work was supported in part by a JSPS Grant-in-Aid for Scientific Research on Innovative Areas (Multidisciplinary Computational Anatomy, #26108005 and #17H05301), JAPAN.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.

Informed consent

Informed consent was obtained from all individuals included in the study.


  1. 1.
    Beriman L (2001) Random forests. Mach Learn 45:5–32CrossRefGoogle Scholar
  2. 2.
    Bresnahan L, Smith J, Ogden A, Quinn S, Cybulski G, Simonian N, Natarajan R, Fessler R, Fessler R (2017) Assessment of paraspinal muscle cross-sectional area after lumbar decompression: minimally invasive versus open approaches. Clin Spine Surg 30(3):E162–E168CrossRefPubMedGoogle Scholar
  3. 3.
    Cooper R, Clair Forbes W, Jayson M (1992) Radiographic demonstration of paraspinal muscle wasting in patients with chronic low back pain. Rheumatology 31(6):389–394CrossRefGoogle Scholar
  4. 4.
    Dubuisson M, Jain A (1994) A modified hausdorff distance for object matching. In: Proceedings of international conference on pattern recognition (ICPR). pp 566–568Google Scholar
  5. 5.
    Engstrom C, Fripp J, Jurcak V, Walker D, Salvado O, Crozier S (2011) Segmentation of the quadratus lumborum muscle using statistical shape modeling. J Magn Reson Imaging 33:1422–1429CrossRefPubMedGoogle Scholar
  6. 6.
    Hides J, Stokes M, Saide M, Jull G, Cooper D (1994) Evidence of lumbar multifidus muscle wasting ipsilateral to symptoms in patients with acute/subacute low back pain. Spine 19(2):165–172CrossRefPubMedGoogle Scholar
  7. 7.
    Inoue T, Kitamura Y, Li Y, Ito W, Ishikawa H (2015) Psoas major muscle segmentation using higher-order shape prior. In: Proceedings of MICCAI-MCV workshop. pp 116–124CrossRefGoogle Scholar
  8. 8.
    Kalichman L, Carmeli E, Been E (2017) The association between imaging parameters of the paraspinal muscles, spinal degeneration, and low back pain. Biomed Res Int 2017:14CrossRefGoogle Scholar
  9. 9.
    Kamiya N, Zhou X, Chen H, Hara T, Hoshi H, Yokoyama R, Kanematsu M, Fujita H (2009) Automated recognition of the psoas major muscles on X-ray CT images. In: Proceedings of IEEE-EMBC 2009. pp 3557–3560Google Scholar
  10. 10.
    Kamiya N, Zhou X, Chen H, Muramatsu C, Hara T, Yokoyama R, Kanematsu M, Hoshi H, Fujita H (2011) Automated segmentation of recuts abdominis muscle using shape model in X-ray CT images. In: Proceedings of IEEE-EMBC 2011. pp 7993–7996Google Scholar
  11. 11.
    Karlsson A, Rosander J, Romu T, Tallberg J, Groenqvist A, Borga M, Dahlqvist Leinhard O (2015) Automatic and quantitative assessment of regional muscle volume by multi-atlas segmentation using whole-body water-fat mri. J Magn Reson Imaging 41(6):1558–1569CrossRefPubMedGoogle Scholar
  12. 12.
    Kume M, Kamiya N, Zhou X, Kato H, Chen H, Muramatsu C, Hara T, Miyoshi T, Matsuo M, Fujita H (2017) Automated recognition of the erector spinae muscle based on deep CNN at the level of the twelfth thoracic vertebrae in torso CT images. In: Proceedings of the 36th JAMIT annual meetingGoogle Scholar
  13. 13.
    Le Troter A, Foure A, Guye M, Confort-Gouny S, Mattei J, Gondin J, Salort-Campana E, Bendahan D (2016) Volume measurements of individual muscles in human quadriceps femoris using atlas-based segmentation approaches. Magn Reson Mater Phys Biol Med (MAGMA) 29(2):245–257CrossRefGoogle Scholar
  14. 14.
    Long J, Shelhamer E, Darrell T (2015) Fully convolutional networks for semantic segmentation. In: Proceedings of 2015 IEEE conference on computer vision and pattern recognition (CVPR 2015). pp 3431–3440Google Scholar
  15. 15.
    Makrogiannis S, Serai S, Fishbein K, Schreiber C, Ferrucci L, Spencer R (2012) Automated quantification of muscle and fat in the thigh from water-, fat-, and nonsuppressed mr images. J Magn Reson Imaging 35(5):1153–1161CrossRefGoogle Scholar
  16. 16.
    Nimura Y, Deguchi D, Kitasaka T, Mori K, Suenaga Y (2008) Pluto: a common platform for computer-aided diagnosis. Med Imaging Technol 26(3):187–191Google Scholar
  17. 17.
    Ogier A, Sdika M, Foure A, Le Troter A, Bendahan D (2017) Individual muscle segmentation in MR images: A 3D propagation through 2D non-linear registration approaches. In: Proceedings of IEEE-EMBC 2017. pp 317–320Google Scholar
  18. 18.
    Orgiu S, Lafortuna C, Rastelli F, Cadioli M, Falini A, Rizzo G (2016) Automatic muscle and fat segmentation in the thigh from t1-weighted MRI. J Magn Reson Imaging 43(3):601–610CrossRefPubMedGoogle Scholar
  19. 19.
    Ozdemir F, Karani N, Fuernstahl P, Goksel O (2017) Interactive segmentation in MRI for orthopedic surgery planning: bone tissue. Int J Comput Assist Radiol Surg 12(6):1031–1039CrossRefPubMedGoogle Scholar
  20. 20.
    Popuri K, Cobzas D, Esfandiari N, Baracos V, Jaegersand M (2016) Body composition assessment in axial CT images using FEM-based automatic segmentation of skeletal muscle. IEEE Trans Med Imaging 35(2):512–520CrossRefPubMedGoogle Scholar
  21. 21.
    Qian C, Wang L, Gao Y, Yousuf A, Yang X, Oto A, Shen D (2016) In vivo MRI based prostate cancer localization with random forests and auto-context model. Comput Med Imaging Graph 52:44–57CrossRefPubMedPubMedCentralGoogle Scholar
  22. 22.
    Sdika M, Tonson A, Le Fur Y, Cozzone P, Bendahan D (2016) Multi-atlas-based fully automatic segmentation of individual muscles in rat leg. Magn Reson Mater Phys Biol Med (MAGMA) 29(2):223–235CrossRefGoogle Scholar
  23. 23.
    Shelhamer E, Long J, Darrell T (2017) Fully convolutional networks for semantic segmentation. IEEE Trans Pattern Anal Mach Intell 39(4):640–651CrossRefPubMedGoogle Scholar
  24. 24.
    Simonyan K, Zisserman A (2014) Very deep convolutional networks for large-scale image recognition. arXiv:1409.1556
  25. 25.
    Tu Z, Bai X (2010) Auto-context and its application to high-level vision tasks and 3D brain image segmentation. IEEE Trans Pattern Anal Mach Intell 32:1744–1757CrossRefPubMedGoogle Scholar
  26. 26.
    Viola P, Jones M (2001) Rapid object detection using a boosted cascade of simple features. In: Proceedings of 2001 CVPR conference. IEEE pp 511–518Google Scholar
  27. 27.
    Viola P, Jones M (2004) Robust real-time face detection. Int J Comput Vis 57:137–154CrossRefGoogle Scholar
  28. 28.
    Wang C, Teboul O, Michel F, Essafi S, Paragios N (2010) 3D knowledge-based segmentation using pose-invariant higher-order graphs. In: Proceedings of MICCAI 2010. vol Part 3. pp 189–196CrossRefGoogle Scholar
  29. 29.
    Wei Y, Xu B, Tao X, Qu J (2015) Paraspinal muscle segmentation in CT images using a single atlas. In: Proceedings of IEEE international conference on progress in informatics and computing (IPC). pp 211–215Google Scholar
  30. 30.
    Yang Y, Chong M, Tay L, Yew S, Yeo A, Tan C (2016) Automated assessment of thigh composition using machine learning for Dixon magnetic resonance images. Magn Reson Mater Phys Biol Med (MAGMA) 29(5):723–731CrossRefGoogle Scholar

Copyright information

© CARS 2018

Authors and Affiliations

  1. 1.School of Information Science and TechnologyAichi Prefectural UniversityNagakuteJapan
  2. 2.Institute for Surgical Technology and BiomechanicsUniversity of BernBernSwitzerland
  3. 3.Department of Electrical, Electronic and Computer Engineering, Faculty of EngineeringGifu UniversityGifuJapan
  4. 4.Department of Brain and Cognitive EngineeringKorea UniversitySeoulRepublic of Korea

Personalised recommendations