In this work, several machine learning methods were studied and support vector machine (SVM) was chosen as a classifier in order to classify dental caries. We used color images of single extracted tooth and generated a feature vector (FV) from each image. Then SVM was applied to classify FVs of different classes. Since heathy tooth, i.e.; non caries (NC) tooth progresses through early stage caries (ESC) to late stage caries (LSC), a two steps detection scheme using SVM has been proposed. In the proposed detection scheme – LSC vs. (ESC+NC) classification was performed followed by ESC vs. NC classification. Apart from SVM, k-means++ clustering algorithm was also applied as an attempt to separate FVs of the entire dataset into two clusters. While comparing clustering result with LSC vs. (ESC+NC) classification result, we observed that SVM performed better than clustering. However, before performing ESC vs. NC classification, k-means++ clustering algorithm was applied to the FVs of ESCs to select representative samples from ESC class since the number of NC data points was too smaller than that of ESC. The misclassification rate of the best SVM model in test set was 7.63% for LSC vs. (ESC+NC) classification, whereas it was 16.67% for ESC vs. NC classification. To the best of our knowledge, this is the first work on dental caries detection from color images. The detection accuracy inspires us to conclude that, with many data points and more precise FV generation scheme, SVM can be applied for detecting dental caries from color images.

[1] T. Vos et al., “Global, regional, and national incidence, prevalence, and years lived with disability for 328 diseases and injuries for 195 countries, 1990-2016: A systematic analysis for the Global Burden of Disease Study 2016,” Lancet, 2017.
[2] B. E. Boser, I. M. Guyon, and V. N. Vapnik, “A training algorithm for optimal margin classifiers,” 2004.
[3] D. Arthur and S. Vassilvitskii, “K-Means++: the Advantages of Careful Seeding,” in Proc ACM-SIAM symposium on discrete algorithms., 2007.
[4] G. V. Trunk, “A Problem of Dimensionality: A Simple Example,” IEEE Trans. Pattern Anal. Mach. Intell., 1979.
[5] S. Shalev-Shwartz and S. Ben-David, Understanding machine learning: From theory to algorithms. 2013.
[6] Y. S. Abu-Mostafa, M. Magdon-Ismail, and H. T. Lin, Learning from Data: A Short Course. 2012.
[7] R. Courant and D. Hilbert, Methods of Mathematical Physics. 2008.
[8] T. Evgeniou, M. Pontil, and T. Poggio, “Regularization Networks and Support Vector Machines,” Adv. Comput. Math., 2000.
[9] R. Izmailov, V. Vapnik, and A. Vashist, “Multidimensional splines with infinite number of knots as SVM kernels,” in Proceedings of the International Joint Conference on Neural Networks, 2013.
[10] F. Girosi, “An Equivalence between Sparse Approximation and Support Vector Machines,” Neural Comput., 1998.
[11] T. Poggio and F. Girosi, “Networks for Approximation arid Learning,” Proc. IEEE, 1990.
[12] D. P. Kingma and J. L. Ba, “Adam: A method for stochastic gradient descent,” ICLR Int. Conf. Learn. Represent., 2015.
[13] J. Duchi, E. Hazan, and Y. Singer, “Adaptive Subgradient Methods for Online Learning and Stochastic Optimization,” JMLR, 2011.
[14] I. Sutskever, J. Martens, G. E. Dahl, and G. E. Hinton, “On the importance of initialization and momentum in deep learning,” Int. Conf. Mach. Learn. 2013, 2013.
[15] A. C. Wilson, R. Roelofs, M. Stern, N. Srebro, and B. Recht, “The Marginal Value of Adaptive Gradient Methods in Machine Learning arXiv : 1705 . 08292v2 [ stat . ML ] 22 May 2018,” in 31st Conference on Neural Information Processing Systems (NIPS, 2017.
[16] K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2016.
[17] S. Ren, K. He, R. Girshick, and J. Sun, “Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks.,” IEEE Trans. Pattern Anal. Mach. Intell., 2017.
[18] R. Caruana and S. Lawrence, “Overfitting in Neural Nets: Backpropagation, Conjugate Gradient, and Early Stopping,” Adv. neural Inf. Process. Syst., 2001.
[19] D. G. Luenberger, Linear and Nonlinear Programming: Second Edition. 2003.
[20] B. Dikmen, “ICDAS II CRITERIA (INTERNATIONAL CARIES DETECTION AND ASSESSMENT SYSTEM),” J. Istanbul Univ. Fac. Dent., 2015.
[21] G. Bradski, “The OpenCV Library,” Dr Dobbs J. Softw. Tools, 2000.
[22] F. Pedregosa et al., “Scikit-learn: Machine Learning in Python Gaël Varoquaux Bertrand Thirion Vincent Dubourg Alexandre Passos PEDREGOSA, VAROQUAUX, GRAMFORT ET AL. Matthieu Perrot,” J. Mach. Learn. Res., 2011.
[23] F. A. Farooqi, A. Khabeer, I. A. Moheet, S. Q. Khan, I. Farooq, and A. S. Arrejaie, “Prevalence of dental caries in primary and permanent teeth and its relation with tooth brushing habits among schoolchildren in Eastern Saudi Arabia,” Saudi Med. J., 2015.