Lev V. Utkin *
Anna A. Meldo
Viktor S. Kryshtapovich
Viktor A. Tiulpin
Ernest M. Kasimov
Maxim S. Kovalev
Received 15 July 2019
A task of the efficient classification system construction for segmented lung nodules as the most important element of an intelligent system for the lung cancer diagnosing based on computed tomography scans is studied. The system is trained on open data sets for the lung cancer diagnosis and solves the problem of classifying the lung tumors detected by a segmentation system, and the problem of reducing the number of false-positive cases. Results of its functioning are the probabilities of various diagnoses. The classification system is based on the following approaches. First, segmented nodules are represented in the form of five histograms that characterize the internal structure of nodules, their surface and their surroundings, which allow us to significantly reduce the data dimensionality and to simplify the classification process. Secondly, Siamese neural networks are used for classification, which are an efficient tool that allows us to carry out the differential diagnosis of oncological diseases. Third, a three-channel classification system is proposed, where the first channel uses a cascade of random forests to classify histograms, the second channel is a Siamese neural network consisting of simple neural networks trained in pairs of histograms, and the third channel uses a Siamese neural consisting of convolutional neural networks trained on pairs of segmented 3D objects from the computed tomography images. The resulting diagnosis is determined on the basis of a combination of the probabilities of the diagnoses obtained in each channel and on the basis of the classification accuracy computed by using testing examples.
Deep neural network, classification, radiomics, lung cancer, artificial intelligence, computed tomography, Siamese neural network.
This work is supported by the Russian Science Foundation under grant 18-11-00078.
Utkin, L. et al. (2019). Three-channel intelligent neoplasm classification system for the diagnosis of lung cancer. Robotics and Technical Cybernetics, 7(3), pp.196-207.
- Litjens, G., Kooi, T., Bejnordi, B., Setio, A., Ciompi, F., Ghafoorian, M., van der Laak, J., van Ginneken, B. and Sánchez, C. (2017). A survey on deep learning in medical image analy-sis. Medical Image Analysis, 42, pp.60-88.
- Siegel, R., Miller, K. and Jemal, A. (2018). Cancer statistics, 2018. CA: A Cancer Journal for Clinicians, 68(1), pp.7-30.
- Zhang, G., Jiang, S., Yang, Z., Gong, L., Ma, X., Zhou, Z., Bao, C. and Liu, Q. (2018). Automatic nodule detection for lung cancer in CT images: A review. Computers in Biology and Medicine, 103, pp.287-300.
- Zhang, J., Xia, Y., Cui, H. and Zhang, Y. (2018). Pulmonary nodule detection in medical images: A survey. Biomedical Signal Processing and Control, 43, pp.138-147.
- Ronneberger, O., Fischer, P. and Brox, T. (2015). U-Net: Convolutional Networks for Biomedical Image Segmentation. Lecture Notes in Computer Science, pp.234-241.
- Çiçek, Ö., Abdulkadir, A., Lienkamp, S., Brox, T. and Ronneberger, O. (2016). 3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation. Medical Image Computing and Computer-Assisted Intervention – MICCAI 2016, pp.424-432.
- Meldo, A., Utkin, L. and Ryabinin, M. (2019). A combined automated diagnostic system for segmentation and detection of lung cancer nodules. Robotics and Technical Cybernetics, 7(2), pp.145-153. (in Russian).
- Chen, B., Zhang, R., Gan, Y., Yang, L. and Li, W. (2017). Development and clinical appli-cation of radiomics in lung cancer. Radiation Oncology, 12(1).
- Thawani, R., McLane, M., Beig, N., Ghose, S., Prasanna, P., Velcheti, V. and Madabhushi, A. (2018). Radiomics and radiogenomics in lung cancer: A review for the clinician. Lung Cancer, 115, pp.34-41.
- Afshar, P., Mohammadi, A., Plataniotis, K., Oikonomou, A. and Benali, H. (2018). From Handcrafted to Deep-Learning-Based Cancer Radiomics: Challenges and opportunities. IEEE Signal Processing Magazine, 36(4), pp.132-160.
- Koch, G., Zemel, R. and Salakhutdinov, R. (2015). Siamese neural networks for one-shot image recognition. Proceedings of the 32nd International Conference on Machine Learning, 37, pp.1-8.
- Li Fei-Fei, Fergus, R. and Perona, P. (2006). One-shot learning of object categories. IEEE Transactions on Pattern Analysis and Machine Intelligence, 28(4), pp.594-611.
- Zhou, Z. and Feng, J. (2017). Deep Forest: Towards An Alternative to Deep Neural Networks. Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence.
- Breiman, L. (2001). Machine Learning, 45(1), pp.5-32.
- Setio, A., Traverso, A., de Bel, T., Berens, M., Bogaard, C., Cerello, P., Chen, H., Dou, Q., Fantacci, M., Geurts, B., Gugten, R., Heng, P., Jansen, B., de Kaste, M., Kotov, V., Lin, J., Manders, J., Sóñora-Mengana, A., García-Naranjo, J., Papavasileiou, E., Prokop, M., Saletta, M., Schaefer-Prokop, C., Scholten, E., Scholten, L., Snoeren, M., Torres, E., Vandemeulebroucke, J., Walasek, N., Zuidhof, G., Ginneken, B. and Jacobs, C. (2017). Validation, comparison, and combination of algorithms for automatic detection of pulmonary nodules in computed tomography images: The LUNA16 challenge. Medical Image Analysis, 42, pp.1-13.
- Utkin, L., Ryabinin, M., Meldo, A., Lukashin, A. and Prokhorov, I. (2019). Baza dannih computernih tomogramm grudnoy kletki s videlannimi i markirovannimi oblastami pathology legkih – LIRA (Lung Image Resource Annotated). 2019620232. (in Russian).
- Smith, S. and Jain, A. (1982). Chord distributions for shape matching. Computer Graphics and Image Processing, 19(1), pp.93-94.
- Chopra, S., Hadsell, R. and LeCun, Y. (2005). Learning a Similarity Metric Discriminatively, with Application to Face Verification. 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05).
- Bellet, A., Habrard, A. and Sebban, M. (2013). A Survey on Metric Learning for Feature Vectors and Structured Data. ArXiv, abs/1306.6709.
- Le Capitaine, H. (2018). Constraint selection in metric learning. Knowledge-Based Systems, 146, pp.91-103.
- Zheng, L., Duffner, S., Idrissi, K., Garcia, C. and Baskurt, A. (2016). Siamese multi-layer perceptrons for dimensionality reduction and face identification. Multimedia Tools and Applications, 75(9), pp.5055-5073.
- Liu, F., Ting, K., Yu, Y. and Zhou, Z. (2008). Spectrum of Variable-Random Trees. Journal of Artificial Intelligence Research, 32, pp.355-384.
- He, K., Zhang, X., Ren, S. and Sun, J. (2016). Deep Residual Learning for Image Recog-nition. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).
- Holzinger, A., Biemann, C., Pattichis, C. and Kell, D. (2017). What do we need to build explainable AI systems for the medical domain? ArXiv, abs/1712.09923.