Skip to main content
Springer Nature Link
Log in

Statistical model-based segmentation of the proximal femur in digital antero-posterior (AP) pelvic radiographs

  • Original Article
  • Published:
International Journal of Computer Assisted Radiology and Surgery Aims and scope Submit manuscript

Abstract

Purpose

   Segmentation of the proximal femur in digital antero-posterior (AP) pelvic radiographs is required to create a three-dimensional model of the hip joint for use in planning and treatment. However, manually extracting the femoral contour is tedious and prone to subjective bias, while automatic segmentation must accommodate poor image quality, anatomical structure overlap, and femur deformity. A new method was developed for femur segmentation in AP pelvic radiographs.

Methods

   Using manual annotations on 100 AP pelvic radiographs, a statistical shape model (SSM) and a statistical appearance model (SAM) of the femur contour were constructed. The SSM and SAM were used to segment new AP pelvic radiographs with a three-stage approach. At initialization, the mean SSM model is coarsely registered to the femur in the AP radiograph through a scaled rigid registration. Mahalanobis distance defined on the SAM is employed as the search criteria for each annotated suggested landmark location. Dynamic programming was used to eliminate ambiguities. After all landmarks are assigned, a regularized non-rigid registration method deforms the current mean shape of SSM to produce a new segmentation of proximal femur. The second and third stages are iteratively executed to convergence.

Results

   A set of 100 clinical AP pelvic radiographs (not used for training) were evaluated. The mean segmentation error was \(0.96\,\hbox {mm} \pm 0.35\,\hbox {mm}\), requiring \(<\!5\) s per case when implemented with Matlab. The influence of the initialization on segmentation results was tested by six clinicians, demonstrating no significance difference.

Conclusions

   A fast, robust and accurate method for femur segmentation in digital AP pelvic radiographs was developed by combining SSM and SAM with dynamic programming. This method can be extended to segmentation of other bony structures such as the pelvis.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Subscribe and save

Springer+ Basic
$34.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or eBook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10

Similar content being viewed by others

Abbreviations

AP:

Antero-posterior

3D:

Three-dimensional

2D:

Two-dimensional

SSM:

Statistical shape model

SAM:

Statistical appearance model

THA:

Total hip arthroplasty

DHS:

Dynamic hip screw

PFN:

Proximal femur nail

CT:

Computed tomography

MRI:

Magnetic resonance imaging

OA:

Osteoarthritis

FAI:

Femoroacetabular impingements

TVO:

Transtrochanteric valgus osteotomy

PCA:

Principal component analysis

References

  1. Zheng G (2010) Statistical shape model-based reconstruction of a scaled, patient specific surface model of the pelvis from a single standard AP X-ray radiograph. Med Phys 37(4):1424–1439

    Article  PubMed  Google Scholar 

  2. Zheng G (2010) Statistically deformable 2D/3D registration for estimating post-operative cup orientation from a single standard AP X-ray radiograph. Ann Biomed Eng 38(9):2910–2927

    Article  PubMed  Google Scholar 

  3. Wright D, Whyne C, Hardisty M, Kreder HJ, Lubovsky O (2011) Functional and anatomic orientation of the femoral head. Clin Orthop Relat Res 469:2583–2589. doi:10.1007/s11999-010-1754-1

    Article  PubMed Central  PubMed  Google Scholar 

  4. Kellgren JH, Lawrence JS (1957) Radiological assessment of osteoarthritis. Ann Rheum Dis 16:494–501

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  5. Turmezei TD, Poole KES (2011) Computed tomography of subchondral bone and osteophytes in hip osteoarthritis: the shape of things to come? Front Endocrinol 2:97. doi:10.3389/fendo.2011.00097

    Article  Google Scholar 

  6. Charbonnier C, Magnenat-Thalmann N, Becker CD, Hoffmeyer P, Menetrey J (2010) An integrated platform for hip joint osteoarthritis analysis: design, implementation and results. Int J Comput Assist Radiol Surg 5(4):351–358

    Article  PubMed  Google Scholar 

  7. Behiels G, Vandermeulen D, Maes F, Suetens P, Dewaele P (1999) Active shape model-based segmentation of digital X-ray images. In: Taylor C et al (eds) MICCAI 1999, LNCS 1679. Springer, Heidelberg, pp 128–137

    Google Scholar 

  8. Pilgram R, Walch C, Kuhn V, Schubert R, Staudinger R (2008) Proximal femur segmentation in conventional pelvic X ray. Med Phys 35(6):2463–2472

    Article  PubMed  Google Scholar 

  9. Lindner C, Thiagarajah S, Wilkinson JM et al (2012) Accurate fully automatic femur segmentation in pelvic radiographs using regression voting. In: Ayache N et al (eds) MICCAI 2012, Part III, LNCS 7512. Springer, Heidelberg, pp 353–360

    Google Scholar 

  10. Chen Y, Ee X, Leow WK, Howe TS (2005) Automatic extraction of femur contours from hip X-ray images. In: Liu Y et al (eds) ICCV Workshop on CVBIA 2005, LNCS 3765. Springer, Heidelberg, pp 200–209

    Google Scholar 

  11. Boukala N, Favier E, Laget B, Radeva P (2004) Active shape model based segmentation of bone structures in hip radiographs. In: IEEE international conference on industrial technology (ICIT’04), pp 1682–1687. doi:10.1109/ICIT.2004.1490821

  12. Gamage P, Xie SQ, Delmas P, Xu WL (2010) Segmentation of radiographic images under topological constraints: application to the femur. Int J Comput Assist Radiol Surg 5:425–435. doi:10.1007/s11548-009-0399-6

    Article  PubMed  Google Scholar 

  13. Ding F, Leow WK, Howe TS (2007) Automatic segmentation of femur bones in anterior-posterior pelvis X-Ray images. In: Kropatsch WG et al (eds) CAIP 2007, LNCS 4673. Springer, Heidelberg, pp 205–212

  14. Cootes TF, Hill A, Taylor CJ, Haslam J (1994) The use of active shape models for locating structures in medical images. Image Vis Comput 12(6):355–366

    Article  Google Scholar 

  15. Cootes TF (2000) An introduction to active shape models. In: Baldock R, Graham J (eds) Image processing and analysis. Oxford University Press, Oxford, pp 223–248

    Google Scholar 

  16. Xie W, Schumann S, Franke J, Grützner PA, Nolte L-P, Zheng G (2012) Finding deformable shapes by correspondence-free instantiation and registration of statistical shape models. In: Wang F et al (eds) MICCAI Workshop on MLMI 2012, LNCS 7588. Springer, Heidelberg, pp 258–265

    Google Scholar 

  17. Xie W, Nolte LP, Zheng G (2011) ECM versus ICP for point registration. In: 33rd Annual international conference of the IEEE engineering in medicine and biology society (EMBC’11), pp 2131–2135. doi:10.1109/IEMBS.2011.6090398

  18. Bellman R (1957) Dynamic programming. Princeton University Press, Dover paperback edition (2003), ISBN 0-486-42809-5

  19. Zheng G, Gollmer S, Schumann S, Dong X, Feilkas T, Ballester MAG (2009) A 2D/3D correspondence building method for reconstruction of a patient-specific 3D bone surface model using point distribution models and calibrated X-ray images. Med Image Anal 13(6):883–899

    Google Scholar 

  20. Zheng G, von Recum J, Nolte L-P, Gruetzner PA, Steppacher SD, Franke J (2012) Validation of a statistical shape model-based 2D/3D reconstruction method for determination of cup orientation after THA. Int J Comput Assist Radiol Surg 7(2):225–231

    Article  CAS  PubMed  Google Scholar 

  21. Jingushi S, Sugioka Y, Noguchi Y, Miura H, Iwamoto Y (2002) Transtrochanteric valgus osteotomy for the treatment of osteoarthritis of the hip secondary to acetabular dysplasia. J Bone Joint Surg Br 84-B(4):535–539

    Google Scholar 

  22. Li W, Kornak J, Harris T et al (2009) Identify fracture-critical regions inside the proximal femur using statistical parametric mapping. Bone 44(4):596–602. doi:10.1016/j.bone.2008.12.008

    Article  PubMed Central  PubMed  Google Scholar 

Download references

Acknowledgments

This works was partially supported by the Swiss National Science foundation via project No. 205321_138009.

Conflict of interest

Weiguo Xie, Jochen Franke, Cheng Chen, Paul A. Grützner, Steffen Schumann, Lutz-P. Nolte, and Guoyan Zheng declare that they have no conflict of interest.

Author information

Authors and Affiliations

  1. Institute for Surgical Technology and Biomechanics, University of Bern, Stauffacherstrasse 78, 3014 , Bern, Switzerland

    Weiguo Xie, Cheng Chen, Steffen Schumann, Lutz-P. Nolte & Guoyan Zheng

  2. Graduate School for Cellular and Biomedical Sciences, University of Bern, Bern, Switzerland

    Weiguo Xie

  3. BG Trauma Centre Ludwigshafen at Heidelberg University Hospital, Ludwigshafen, Germany

    Weiguo Xie, Jochen Franke & Paul A. Grützner

Authors
  1. Weiguo Xie
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Jochen Franke
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Cheng Chen
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Paul A. Grützner
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Steffen Schumann
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Lutz-P. Nolte
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Guoyan Zheng
    View author publications

    You can also search for this author inPubMed Google Scholar

Corresponding author

Correspondence to Weiguo Xie.

Electronic supplementary material

Below is the link to the electronic supplementary material.

11548_2013_932_MOESM1_ESM.bmp

On-line supplemental Fig. I: AP pelvic radiographs clinically acquired for different applications: a) dynamic hip screw (DHS); b) proximal femur nail (PFN); c) total hip arthroplasty (THA) (post-operative); d) THA (pre-operative). (bmp 1761 KB)

11548_2013_932_MOESM2_ESM.bmp

On-line supplemental Fig. II: a) Five manually acquired points (red dots) for initialization of the algorithm and the mean shape (green stars and dots) of the SSM before initialization. b) The roughly registered mean shape (green stars and dots) of the SSM after initialization. c) First iteration of the SAM-based registration stage: suggested points (blue stars) found for all landmarks on the mean shape of the SSM. d) First iteration of the SSM-based deformation stage: the deformed mean shape (yellow dots and stars) of the SSM. e) Final segmentation results (yellow dots and stars) after three complete iterations of both SAM- and SSM-based registration stages. f) Comparison: the continuous contour (yellow) obtained by a b-spline interpolation of the segmentation results and the ground truth contour obtained by manual segmentation (green). (bmp 2631 KB)

Rights and permissions

Reprints and permissions

About this article

Cite this article

Xie, W., Franke, J., Chen, C. et al. Statistical model-based segmentation of the proximal femur in digital antero-posterior (AP) pelvic radiographs. Int J CARS 9, 165–176 (2014). https://doi.org/10.1007/s11548-013-0932-5

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11548-013-0932-5

Keywords