Brain Tumor MRI Image Segmentation and Classification based on Deep Learning Techniques

Abstract

Detection and diagnosis of brain tumors is important for improving the possibility of successful treatment and recovering. Magnetic resonance imaging (MRI) is widely used imaging method for treating and recovering brain tumors. However, manual identification of brain tumors from a large amount of MRI images is time-consuming and requires specialized expertise. To overcome these challenges, computer- assisted intelligent systems are increasingly being used to speed up the medical assessment as well as treatment recommendations. The aim of our research is for coming up with a deep learning system that can segment and classify tumors in brain. The U-Net model is used for segmentation of the MRI images, while Convolution Neural Network (CNN) is used for the classifying brain tumors. Performance metrics such as accuracy, precision and recall are used to evaluate the effectiveness of this approach. The suggested CNN classifier has given the accuracy of nearly 98% for both the training and validation data. By using deep learning techniques, the following system attempts to provide accurate and effective segmenting and classifying tumors.