Abstract—The formation of abnormal cells within the brain
is termed as brain tumor. Brain tumors can be benign or
cancerous. Brain being the major organ of the body controls
every activity we perform in our daily life. Therefore, it is
important to detect any kind of abnormal growth of cells in the
brain at the earliest. Deep learning methods are now being
hugely used in medical image analysis. Manual detection of
brain tumors by radiologists are time consuming and error
prone. Therefore applying Machine Learning techniques to
automatically segment tumor region and to detect is very
important and necessary in advancing medical image analysis.
The algorithm in this paper could automatically detect brain
tumor through skull stripping method and segmentation
through U-Net architecture. The algorithm has been tested on
3000 Magnetic resonance imaging images (MRI) and resulted
in an accuracy of 93%. Dataset of Digital Imaging and
Communications in Medicine (DICOM) format brain MRI
images has been used for the experimentation. The proposed
method achieved mean Dice Similarity Coefficient metric of
0.82 and median Dice Similarity Coefficient metric of 0.86 for
full tumor region.
Index Terms—Skull stripping, U-Net architecture, data
augmentation.
Debapriya Hazra and Yungcheol Byun are with the Jeju National University,
South Korea (corresponding author: Yungcheol Byun; e-mail:
debapriyah@gmail.com, yungcheolbyun@gmail.com).
Cite: Debapriya Hazra and Yungcheol Byun, "Brain Tumor Detection Using Skull Stripping and U-Net Architecture," International Journal of Machine Learning and Computing vol. 10, no. 2, pp. 400-405, 2020.
Copyright © 2020 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).