Brain Tumors Detection using Computed Tomography Scans Based on Deep Neural Networks

Authors

N. M. Dawood, Physics Department, Faculty of Science, Mansoura University, Mansoura 35516, EgyptFollow
L. M. AbouEl-Magd, Computer Science Department, Misr Higher Institute of Computers, Mansoura, EgyptFollow
A.-H. Abdel-Aty, Physics Department, College of Science, University of Bisha, P.O. Box 344, Bisha 61922, Saudi ArabiaFollow
W. S. Awad, College of Information Technology, Ahlia University, Kingdom of BahrainFollow

Abstract

Brain tumors are one of the deadliest diseases, with numerous implications on human health. A brain tumor is an abnormal cell mass or growth in or around the brain. They are not all cancerous, as they might be benign or malignant. Doctors use a variety of diagnostic techniques to assess the presence of a benign or malignant brain tumor, as well as to estimate its size, location, and growth rate. The proper diagnostic modality is used to provide a complete view of the brain to detect any abnormalities. A computed tomographic (CT) scan of the brain shall be done to check the abnormalities. The benefits of CT scans include accurate detection of calcification, hemorrhage, and bone detail, as well as low cost compared to magnetic resonance imaging (MRI). Therefore, we examine a proposed CT-based detection method to determine whether brain tumor is present or not. The proposed method works on a CT image dataset that collected from Mansoura University hospital. Different pre-trained models are used: VGG-16, ResNet-50, and MobileNet-V2. Comparing the results, that pre-train model MobileNet-V2, despite having the lowest number of parameters, yields better results. It gives an accuracy 97.6%, while its precision, recall, and F1-score values are 96%, 95%, and 96%, respectively.

Recommended Citation

M. Dawood, N.; M. AbouEl-Magd, L.; Abdel-Aty, A.-H.; and S. Awad, W. (2023) "Brain Tumors Detection using Computed Tomography Scans Based on Deep Neural Networks," Information Sciences Letters: Vol. 12 : Iss. 4 , PP -.
Available at: https://digitalcommons.aaru.edu.jo/isl/vol12/iss4/33