Error Estimation for Appearance Model Segmentation of Musculoskeletal Structures Using Multiple, Independent Sub-models
Segmentation of structures in clinical images is a precursor to computer-aided detection (CAD) for many musculoskeletal pathologies. Accurate CAD systems could considerably improve the efficiency and objectivity of radiological practice by providing clinicians with image-based biomarkers calculated with minimal human input. However, such systems rarely achieve human-level performance, so extensive manual checking may be required. Their practical utility could therefore be increased by accurate error estimation, focusing manual input on the images or structures where it is needed. Standard techniques such as the minimum variance bound can estimate random errors, but provide no way to estimate any systematic errors due to model fitting failure.
We describe the use of multiple, independent sub-models to estimate both systematic and random errors. The approach is evaluated on vertebral body segmentation in lateral spinal images, demonstrating large (up to 50%) and significant improvements in the accuracy of error classification with concurrent improvements in annotation accuracy. Whilst further work is required to elucidate the definition of “independence” in this context, we conclude that the approach provides a valuable component for appearance model based CAD systems.
This publication presents independent research supported by the NIHR Invention for Innovation (i4i) programme (grant no. II-LB_0216-20009). The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health and Social Care. The authors acknowledge the invaluable assistance of Mrs Chrissie Alsop, Mr Stephen Capener, Mrs Imelda Hodgkinson, Mr Michael Machin, and Mrs Sue Roberts, who performed the manual annotations.
- 3.Bromiley, P.A., Adams, J.E., Cootes, T.F.: Localisation of vertebrae on DXA images using constrained local models with random forest regression voting. In: Yao, J., Glocker, B., Klinder, T., Li, S. (eds.) Recent Advances in Computational Methods and Clinical Applications for Spine Imaging. LNCVB, vol. 20, pp. 159–171. Springer, Heidelberg (2015). https://doi.org/10.1007/978-3-319-14148-0_14CrossRefGoogle Scholar
- 4.Bromiley, P.A., Kariki, E.P., Adams, J.E., Cootes, T.F.: Fully automatic localisation of vertebrae in CT images using random forest regression voting. In: Yao, J., Vrtovec, T., Zheng, G., Frangi, A., Glocker, B., Li, S. (eds.) CSI 2016. LNCS, vol. 10182, pp. 51–63. Springer, Cham (2016). https://doi.org/10.1007/978-3-319-55050-3_5CrossRefGoogle Scholar
- 5.Bromiley, P.A., Pokric, M., Thacker, N.A.: Empirical evaluation of covariance estimates for mutual information coregistration. In: Barillot, C., Haynor, D.R., Hellier, P. (eds.) MICCAI 2004. LNCS, vol. 3216, pp. 607–614. Springer, Heidelberg (2004). https://doi.org/10.1007/978-3-540-30135-6_74CrossRefGoogle Scholar
- 6.Bromiley, P.A., Schunke, A.C., Ragheb, H., Thacker, N.A., Tautz, D.: Semi-automatic landmark point annotation for geometric morphometrics. Front. Zool. 11(61), 1–21 (2014)Google Scholar
- 9.Erdt, M., Steger, S., Wesarg, S.: Deformable registration of MR images using a hierarchical patch based approach with a normalized metric quality measure. In: 2012 9th IEEE International Symposium on Biomedical Imaging (ISBI), pp. 1347–1350 (2012)Google Scholar
- 11.Jergas, M., Valentin, R.S.: Techniques for the assessment of vertebral dimensions in quantitative morphometry. In: Genant, H.K., Jergas, M., van Juijk, C. (eds.) Vertebral Fracture In Osteoporosis, pp. 163–188. University of California Osteoporosis Research Group, San Francisco (1995)Google Scholar
- 12.Kariki, E.P., Bromiley, P.A., Cootes, T.F., Adams, J.A.: Opportunistic identification of vertebral fractures on computed radiography: need for improvement. Osteoporos. Int. 27(S2), 621 (2016)Google Scholar