Abstract
Brain tumor accounts for 80–90% of all primary CNS tumors. Brain cancer is tenth leading cause of death. Automatic detection and classification can help in early diagnosis and deliver efficient treatment. Recent developments in medical imaging modalities such as MRI provide insightful image of our brain. Due to large amount data and variability of data, diagnosing in faster and proper manner is not humanly possible. In this paper, we focus on segmentation of brain tumor of 3D MRI images using 3D CNN in cascading format. Whole tumor is extracted from the images through the first CNN model, then the output is fed to the next CNN model to extract the core tumor and finally, it is fed to the last CNN of cascading network to segment the enhanced tumor core. To improve the brain tumor segmentation, three neural networks belonging to the class of convoluted neural networks (CNNs) were connected having 20 interconnected kernel slices with four downsampling slices to trade off simplicity with feature extraction. The proposed method was evaluated by considering Brats 2015 3D dataset consisting of 274 MRI images with their ground truth having different four modalities. The activation function used was Relu and the results obtained were calculated over DICE coefficient (F-measure) which was found to be 0.78 for core tumor in flair modalities.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Bakas, S., Akbari, H., Sotiras, A., Bilello, M., Rozycki, M., Kirby J., et al. (2018) Advancing the cancer genome atlas glioma MRI automatic brain tumor segmentation using cascaded anisotropic CNNs.
Razavian, A. S., Azizpour, H., Sullivan, J., & Carlsson, S. (2014) CNN features off-the-shelf: an astounding baseline for recognition.
Milletari, F., Navab, & N., Ahmadi, S. A. (2016). V-Net: Fully convolutional neural networks for volumetric medical image segmentation. In IC3DV. pp. 565–571.
Alqazzaz, S., Sun, X., Yang, X., & Nokes, L. (2019). Automated brain tumor segmentation on multi-modal MR image using SegNet. Computational Visual Media, 1–11.
He, K., Zhang, X., Ren, S., & Sun, J. (2015). Delving deep into rectifiers: surpassing human-level performance on ImageNet classification. In ICCV, pp. 1026–1034.
Long, J., Shelhamer, E., & Darrell, T. (2015). Fully convolutional networks for semantic segmentation. In CVPR, pp. 3431–3440.
Kamnitsas, K., Ledig, C., Newcombe, V. F. J., Simpson, J. P., Kane, A. D., Menon, D. K., et al. (2017). Efficient multi-scale 3D CNN with fully connected CRF for accurate brain lesion segmentation. Medical Image Analysis, 36, 61–78.
Fidon, L., Li, W., & Garcia-peraza Herrera, L. C. (2017). Generalised Wasserstein dice score for imbalanced multi-class segmentation using holistic convolutional networks, arXiv preprint arXiv:1707.00478.
Fred, A., & Agarap, M. (2019). Deep learning using rectified linear units (ReLU). arXiv:1803.08375.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Das, S., Swain, M.K., Nayak, G.K., Saxena, S. (2021). Brain Tumor Segmentation from 3D MRI Slices Using Cascading Convolutional Neural Network. In: Mallick, P.K., Bhoi, A.K., Chae, GS., Kalita, K. (eds) Advances in Electronics, Communication and Computing. ETAEERE 2020. Lecture Notes in Electrical Engineering, vol 709. Springer, Singapore. https://doi.org/10.1007/978-981-15-8752-8_12
Download citation
DOI: https://doi.org/10.1007/978-981-15-8752-8_12
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-15-8751-1
Online ISBN: 978-981-15-8752-8
eBook Packages: Computer ScienceComputer Science (R0)