Automatic brain tumor detection in MRI images using convolutional neural networks

Abstract

Brain tumors are regarded one of the most aggressive disorders in both children and adults. Brain tumors make for 85-90 percent of all primary central nervous system (CNS) malignancies. Every year, around 11700 people are diagnosed with brain tumors. The 5-year survival rate for patients with a malignant brain or CNS tumor is around 34% for males and 36% for women. Brain tumors are divided as benign, malignant, and pituitary tumors. Proper therapy, planning, and precise diagnostics should be implemented to increase patients' life expectancy. Magnetic resonance imaging (MRI) is the most effective tool for detecting brain malignancies. The scans generate a vast amount of picture data. The radiologist examines these photographs. Because of the intricacies and features of brain tumors, a manual examination can be prone to errors. Automated classification techniques based on Machine Learning (ML) and deep learning have consistently outperformed manual classification. The dataset we have used for this work contains 2 folders: yes and no which contains 3060 brain MRI images. After implementation our custom CNN model achieved 98% accuracy. As a result, offering a system for detection and classification that uses deep learning algorithms such as convolution-neural network (CNN) will be beneficial to doctors all over the world.