Brain tumor segmentation with simplified U-Net architecture

Yıl 2022, , 856 - 861, 14.10.2022

Özlem Polat

https://doi.org/10.28948/ngumuh.1111082

Öz

Brain tumors located in the skull are among the health problems that cause serious consequences. Rapid and accurate detection of the brain tumor and segmentation of the tumor region will increase the patient's chance of recovery and survival by ensuring that the patient receives appropriate treatment in the early period. There are many segmentation methods in the literature. The low segmentation accuracy and the very large network structure used are the main disadvantages of the existing methods. In this study, a simplified U-Net deep learning model is used for segmentation on MR images of brain tumors. The model was trained and tested on 3064 MR images from 233 patients, which included the common brain tumors glioma, meningioma, and pituitary. As a result, average 0.86 dice similarity coefficient, 0.76 IoU score, 0.85 sensitivity value and 0.99 pixel accuracy value were obtained. Since the proposed model performs brain tumor segmentation quickly and with high accuracy, it promises to help specialists in the diagnosis of the disease and in determining the appropriate treatment.

Anahtar Kelimeler

Brain tumor, Segmentation, Deep learning, U-Net

Sadeleştirilmiş U-Net mimarisi ile beyin tümörü segmentasyonu

Öz

Kafatası içinde yer alan beyin tümörleri ciddi sonuçlara neden olan sağlık sorunları arasındadır. Beyin tümörünün hızlı ve doğru bir şekilde tespit edilip tümör bölgesinin segmentasyonunun yapılması hastanın erken dönemde uygun tedavi almasını sağlayarak hastanın iyileşme ve hayatta kalma şansını artıracaktır. Literatürde birçok segmentasyon yöntemi bulunmaktadır. Düşük segmentasyon doğruluğu ve kullanılan ağ yapısının çok büyük olması mevcut yöntemlerin ana dezavantajıdır. Bu çalışmada beyin tümörlerinin MR görüntüleri üzerinde segmentasyonu için sadeleştirilmiş U-Net derin öğrenme modeli önerilmektedir. Model, 233 hastadan alınan ve yaygın beyin tümörlerinden gliom, menenjiom ve hipfiz tümörünü içeren 3064 MR görüntüsü üzerinde eğitilip test edilmiştir. Sonuç olarak ortalama 0.86 dice benzerlik katsayısı, 0.76 IoU skoru, 0.85 hassasiyet değeri ve 0.99 piksel doğruluk değeri elde edilmiştir. Önerilen model beyin tümörü segmentasyonunu hızlı ve yüksek doğrulukla gerçekleştirdiğinden hastalığın tanısında ve uygun tedavinin belirlenmesinde uzmanlara yardımcı olması açısından umut vaat etmektedir.

Anahtar Kelimeler

Beyin tümörü, Segmentasyon, Derin öğrenme, U-Net

Kaynakça

• H. Mohsen, E.A. El-Dahshan, E.M. El-Horbaty ve A.M. Salem, Classification using deep learning neural networks for brain tumors. Future Computing and Informatics Journal, 3 (1), 68-71, 2018. https://doi.org/10.1016/j.fcij.2017.12.001.
• G.S. Tandel, A. Balestrieri, T. Jujaray, N.N. Khanna, L. Saba ve J.S. Suri, Multiclass magnetic resonance imaging brain tumor classification using artificial intelligence paradigm. Computers in Biology and Medicine, 122, 103804, 2020. https://doi.org/10.1016/j.compbiomed.2020.103804.
• Types of cancer, Brain Tumor: Statistics. https://www.cancer.net/cancer-types/brain-tumor/statistics, Accessed 28 April 2022.
• K. Usman ve K. Rajpoot, Brain tumor classification from multi-modality MRI using wavelets and machine learning. Pattern Analysis and Applications, 20, 871-881, 2017. https://doi.org/10.1007/s10044-017-0597-8
• N. Gordillo, E. Montseny ve P. Sobrevilla, State of the art survey on MRI brain tumor segmentation. Magnetic Resonance Imaging, 31 (8), 1426-1438, 2013. https://doi.org/10.1016/j.mri.2013.05.002.
• A. Baştuğ Koç ve D. Akgün, U-net mimarileri ile glioma tümör segmentasyonu üzerine bir literatür çalışması. Avrupa Bilim ve Teknoloji Dergisi, 26, 407-414, 2021. https://doi.org/10.31590/ejosat.959590.
• Ö. Polat, C. Güngen, Classification of brain tumors from MR images using deep transfer learning. Journal of Supercomputing, 77, 7236–7252, 2021. https://doi.org/10.1007/s11227-020-03572-9.
• G. Xiao, H. Wang, J. Shen, Z. Chen, Z. Zhang ve X. Ge, Synergy factorized bilinear network with a dual suppression strategy for brain tumor classification in MRI. Micromachines, 13 (1), 2022. https://doi.org/10.3390/mi13010015.
• B.V. Isunuri ve J. Kakarla, Three-class brain tumor classification from magnetic resonance images using separable convolution based neural network. Concurrency and Computation: Practice and Experience, 34, e6541, 2022. https://doi.org/10.1002/cpe.6541.
• J. Kakarla, B.V. Isunuri, K.S. Doppalapudi ve K.S.R. Bylapudi, Three-class classification of brain magnetic resonance images using average-pooling convolutional neural network. International Journal of Imaging Systems and Technology, 31 (3), 1731–1740, 2021. https://doi.org/10.1002/ima.22554.
• Z. Sobhaninia, S. Rezaei, N. Karimi, A. Emami ve S. Samavi, Brain Tumor Segmentation by Cascaded Deep Neural Networks Using Multiple Image Scales. 28th Iranian Conference on Electrical Engineering (ICEE), sayfa 1-4, Tabriz, Iran, 4-6 Ağustos 2020. https://doi.org/10.1109/ICEE50131.2020.9260876.
• H.N.T.K. Kaldera, S.R. Gunasekara ve M.B. Dissanayake, MRI based glioma segmentation using deep learning algorithms. 2019 International Research Conference on Smart Computing and Systems Engineering (SCSE), sayfa 51-56, Colombo, Sri Lanka, 28-28 Mart 2019. https://doi.org/10.23919/SCSE.2019.8842668.
• B. Maas, E. Zabeh ve S. Arabshahi, QuickTumorNet: Fast automatic multi-class segmentation of brain tumors. 10th International IEEE/EMBS Conference on Neural Engineering (NER), sayfa 81-85, Italy, 4-6 Mayıs 2021. https://doi.org/10.1109/NER49283.2021.9441286.
• F.J. Díaz-Pernas, M. Martínez-Zarzuela, M. Antón-Rodríguez ve D. González-Ortega, A deep learning approach for brain tumor classification and segmentation using a multiscale convolutional neural network. Healthcare (Basel), 9 (2), 153, 2021. https://doi.org/10.3390/healthcare9020153.
• K. Kamnitsas, C. Ledig, V.F.J. Newcombe, J.P. Simpson, A.D. Kane, D.K. Menon, D. Rueckert ve B. Glocker, Efficient multi-scale 3D CNN with fully connected CRF for accurate brain lesion segmentation. Medical Image Analysis, 36, 61-78, 2017. https://doi.org/10.1016/j.media.2016.10.004.
• J. Cheng, Brain tumor dataset, figshare. 2017. https://doi.org/10.6084/m9.figshare.1512427.v5.
• H. Li, A. Li and M. Wang, A novel end-to-end brain tumor segmentation method using improved fully convolutional networks. Computers in Biology and Medicine, 108, 150-160, 2019. https://doi.org/10.1016/j.compbiomed.2019.03.014.
• D. Daimary, M.B. Bora, K. Amitab ve D. Kandar, Brain tumor segmentation from MRI images using hybrid convolutional neural networks. Procedia Computer Science, 167, 2419-2428, 2020. https://doi.org/10.1016/j.procs.2020.03.295.
• W. Wu, D. Li, J. Du, X. Gao, W. Gu, F. Zhao, X. Feng and H. Yan, An intelligent diagnosis method of brain MRI tumor segmentation using deep convolutional neural network and SVM algorithm. Computational and Mathematical Methods in Medicine, 6789306, 2020. https://doi.org/10.1155/2020/6789306.
• J.L. Foo, A survey of user interaction and automation in medical image segmentation methods. Iowa State University Human Computer Interaction Technical Report ISU-HCI-2006-02, 2006.
• N. Gordillo, E. Montseny ve P.Sobrevilla, State of the art survey on MRI brain tumor segmentation. Magnetic Resonance Imaging, 31 (8), 1426-1438, 2013. https://doi.org/10.1016/j.mri.2013.05.002.
• O. Ronneberger, P. Fischer, ve T. Brox, U-Net: Convolutional Networks for Biomedical Image Segmentation. In: Navab, N., Hornegger, J., Wells, W., Frangi, A. (eds) Medical Image Computing and Computer-Assisted Intervention – MICCAI 2015. MICCAI 2015. Lecture Notes in Computer Science, 9351. Springer, Cham. https://doi.org/10.1007/978-3-319-24574-4_28.
• J. Liang, X. Lv, C. Lu, X. Ye, X. Chen, J. Fu, C. Luo ve Y. Zhao, Prognostic factors of patients with Gliomas - an analysis on 335 patients with glioblastoma and other forms of gliomas. BMC Cancer, 20 (1), 35, 2020. https://doi.org/10.1186/s12885-019-6511-6.
• N. Sarbu, L. Oleaga, I. Valduvieco, T. Pujol ve J. Berenguer, Increased signal intensity in FLAIR sequences in the resection cavity can predict progression and progression-free survival in gliomas. Neurocirugia (Astur), 27 (6), 269-276, 2016. https://doi.org/10.1016/j.neucir.2016.04.002.