Impact of Ensemble Learning in the Assessment of Skeletal Maturity
- 186 Downloads
The assessment of the bone age, or skeletal maturity, is an important task in pediatrics that measures the degree of maturation of children’s bones. Nowadays, there is no standard clinical procedure for assessing bone age and the most widely used approaches are the Greulich and Pyle and the Tanner and Whitehouse methods. Computer methods have been proposed to automatize the process; however, there is a lack of exploration about how to combine the features of the different parts of the hand, and how to take advantage of ensemble techniques for this purpose. This paper presents a study where the use of ensemble techniques for improving bone age assessment is evaluated. A new computer method was developed that extracts descriptors for each joint of each finger, which are then combined using different ensemble schemes for obtaining a final bone age value. Three popular ensemble schemes are explored in this study: bagging, stacking and voting. Best results were achieved by bagging with a rule-based regression (M5P), scoring a mean absolute error of 10.16 months. Results show that ensemble techniques improve the prediction performance of most of the evaluated regression algorithms, always achieving best or comparable to best results. Therefore, the success of the ensemble methods allow us to conclude that their use may improve computer-based bone age assessment, offering a scalable option for utilizing multiple regions of interest and combining their output.
KeywordsBone age Ensembles Regression Bagging Digital imaging Machine learning
This work was partially supported by the Cloud Thinking project (CENTRO-07-ST24-FEDER-002031), co-funded by “Quadro de Referência Estratégica Nacional “(QREN), “Mais Centro” program.
Conflict of Interest
The authors declare that they have no conflict of interest.
- 1.Poznanski, A. K., Hernandez, R. J., Guire, K. E., Bereza, U. L., and Garn, S. M., Carpal length in children–a useful measurement in the diagnosis of rheumatoid arthritis and some concenital malformation syndromes. Radiology 129(3):661–668, 1978.Google Scholar
- 2.Greulich, W. W., and Pyle, S. I., Radiographic Atlas of Skeletal Development of Hand and Wrist, 2nd edition. Stanford Univ. Press, Stanford, 1959.Google Scholar
- 3.Tanner, J. M., Cameron, N., Goldstein, H., and Healy, M. J. R., Assessment of skeletal maturity and prediction of adult height (TW3 method). Saunders Co., WB, 2001.Google Scholar
- 4.Berst, M. J., Dolan, L., Bogdanowicz, M. M., Stevens, M. A., Chow, S., and Brandser, E. A., Effect of knowledge of chronologic age on the variability of pediatric bone age determined using the Greulich and Pyle standards. AJR American journal of roentgenology 176(2):507–510, 2001. doi: 10.2214/ajr.176.2.1760507.CrossRefGoogle Scholar
- 6.King, D. G., Steventon, D. M., O’Sullivan, M. P., Cook, A. M., Hornsby, V. P., Jefferson, I. G., and King, P. R., Reproducibility of bone ages when performed by radiology registrars: an audit of Tanner and Whitehouse II versus Greulich and Pyle methods. The British journal of radiology 67(801):848–851, 1994.CrossRefGoogle Scholar
- 12.Niemeijer M, van Ginneken B, Maas CA, Beek FJ, Viergever MA Assessing the skeletal age from a hand radiograph: automating the Tanner- Whitehouse method. In, 2003. pp 1197–1205Google Scholar
- 14.Thodberg, H. H., and Sävendahl, L., Validation and Reference Values of Automated Bone Age Determination for Four Ethnicities. Academic Radiology 17(11):1425–1432, 2010. doi: 10.1016/j.acra.2010.06.007.
- 15.Cootes, T. F., Taylor, C. J., Cooper, D. H., and Graham, J., Active shape models-their training and application. Computer vision and image understanding 61(1):38–59, 1995.Google Scholar
- 16.Kroon D-J (2010) Active Shape Model (ASM) and Active Appearance Model (AAM): Cootes 2D/3D Active Shape & Appearance Model for automatic image object segmentation and recognition. http://www.mathworks.com/matlabcentral/fileexchange/26706-active-shape-model-asm–and-active-appearance-model–aam-.
- 17.Kroon D-J (2011) Segmentation of the mandibular canal in cone-beam CT data. Enschede, the NetherlandsGoogle Scholar
- 21.Zuiderveld K (1994) Contrast limited adaptive histogram equalization. In: Paul SH (ed) Graphics gems IV. Academic Press Professional, Inc., pp 474–485Google Scholar
- 26.Kuncheva LI (2004) Combining Pattern Classifiers: Methods and Algorithms. Wiley-Interscience,Google Scholar