2518-1092Research result. Information technologies2518-109210.18413/2518-1092-2019-4-2-0-41701SYSTEM ANALYSIS AND PROCESSING OF KNOWLEDGETHE DECISION OF THE TASK OF CLASSIFICATION OF HUMAN BRAIN PATHOLOGIES ON MRI IMAGESTHE DECISION OF THE TASK OF CLASSIFICATION OF HUMAN BRAIN PATHOLOGIES ON MRI IMAGESMikhelevVladimir MikhailovichMikhelevVladimir Mikhailovichvm.mikhelev@gmail.comMiroshnichenkoAndrey SergeyevichMiroshnichenkoAndrey Sergeyevich963565@bsu.edu.ru20194200

The article discusses the methods of classification of MRI images of the human brain. The methods are trained binary classifiers to determine the presence or absence of a high stage of brain pathologies. The study aims to identify the best classification method on the processed BRATS data set. We analyzed three classification methods: image classification by basic contour primitives, a method based on a convolutional neural network, with a binary classifier, and a classification method based on a convolutional pretrained neural network Xception. The paper shows that the use of pre-trained neural convolutional networks allows to reduce the time for calculating the time and resources of a computer system for training a neural network. The results of a computational experiment are presented and it is shown that the best accuracy in solving the problem of pathology classification in MRI scans of the human brain is achieved using the Xception neural convolutional pre-trained neural network.

The article discusses the methods of classification of MRI images of the human brain. The methods are trained binary classifiers to determine the presence or absence of a high stage of brain pathologies. The study aims to identify the best classification method on the processed BRATS data set. We analyzed three classification methods: image classification by basic contour primitives, a method based on a convolutional neural network, with a binary classifier, and a classification method based on a convolutional pretrained neural network Xception. The paper shows that the use of pre-trained neural convolutional networks allows to reduce the time for calculating the time and resources of a computer system for training a neural network. The results of a computational experiment are presented and it is shown that the best accuracy in solving the problem of pathology classification in MRI scans of the human brain is achieved using the Xception neural convolutional pre-trained neural network.

image classification methodsMRI scansneural networksconvolutional networksimage classification methodsMRI scansneural networksconvolutional networksСписок литературыWorld Health Organization, International Agency for Research on Cancer / PRESS-RELEASE No. 223 / [Electronic resource]. – Access mode: http://globocan.iarc.fr/Default.aspx.Miroshnichenko A.S., Mikhelev V.M., Konyaeva E.S. – “Method of image classification” // XIX International Conference “Computer Science: Problems, Methodology, Technologies” (IPMT-2019) and X School-Conference “Computer Science in Education” (INED-2019), 14-15 February, Voronezh.Miroshnichenko A.S., Mikhelev V.M. – “The method of recognition of objects in MRI images based on a convolutional neural network” // XVIII International Conference “Computer Science: problems, methodology, technologies” (IPMT-2018) and IXED – 2018, P. 181-185, 8-9 February, Voronezh.MRI scan of the human brain [electronic resource]. – Access mode: http://eegeasy.com/articles/detail.php?ELEMENT_ID=26.Modern types of tomography. Tutorial / M.Ya. Marusina, A.O. Kaznacheeva. – SPb .: SPbSU ITMO., 2006. – 152 p.Francois Chollet, Xception: Deep Learning with Depthwise Separable Convolutions.2017, arXiv:1610.02357v3 [cs.CV].K. He, X. Zhang, S. Ren, and J. Sun. Deep residual learning for image recognition. 2015, arXiv:1512.03385v1 [cs.CV].K. He, X. Zhang, S. Ren, and J. Sun. Delving deep into rectifiers: Surpassing human-level performance on imagenet classification. 2015, arXiv:1502.01852v1 [cs.CV].S. Ioffe and C. Szegedy. Batch normalization: Accelerating deep network training by reducing internal covariate shift. 2015, arXiv:1502.03167v3 [cs. LG].Krizhevsky, I. Sutskever, and G. E. Hinton. Imagenet classification with deep convolutional neural networks. In NIPS, 2012.Min Lin, Qiang Chen, and Shuicheng Yan. Network in network. CoRR, abs/1312.4400, 2013.Manoj, L.L. Brain Tumor Detection and Segmentation Using Histogram Thresholding/ K.K. Manoj,Y.K. Soubrabh// International Journal of Engineering and Advanced Technology. – 2012. – Vol. 1. – Issue 04. – P. 16-20.Mansfield, P. NMR Imaging in Biomedicine: Supplement 2 Advances in Magnetic Resonance/ P. Mansfield, P.G. Morris. – New York: Academic press, 1982. – 365 P.Neeraj, S. Automated medical image segmentation techniques/ S. Neeraj, L.M. Aggarwal // Journal of medical physics. – 2010. – Vol. 35. – PP. 3-14.Sanjeev Arora, Aditya Bhaskara, Rong Ge, and Tengyu Ma. Provable bounds for learning some deep representations. CoRR, abs/1310.6343, 2013.K. Simonyan and A. Zisserman. Very deep convolutional networks for large-scale image recog-nition. 2015, arXiv:1409.1556v6 [cs.CV].C. Szegedy, V. Vanhoucke, S. Ioffe, J. Shlens, and Z. Wojna. Rethinking the inception architecture for computer vision. arXiv preprint arXiv:1512.00567, 2015.C. Szegedy, W. Liu, Y. Jia, et al. Going deeper with convolutions. In 2015 IEEE Conference on Computer Vision and Pattern Recognition, pages 1–9, 2015.Swe, Z.O. Brain tumor detection and segmentation using watershed segmentation and morphological operation/ Z.O. Swe, S.K. Aung // International Journal of Research in Engineering and Technology. – 2014. – Vol. 3. – Issue 03. – PP. 367-374.M.D. Zeiler, G.W. Taylor, and R. Fergus. Adaptive deconvolutional networks for mid and high-level feature learning. In ICCV ’11 Proceedings of the 2011 International Conference on Computer Vision, pages 2018–2025, 2011.