Brain Tumor Detection Using Deep Learning : Design of Neural Architectures and Deployment of a Graphical Interface

Abstract

Early detection of brain tumors is essential to improve patients' chances of survival. To this end, automatic detection methods are continually being developed. Diagnosis of this serious disease can be lengthy and vary between doctors. Deep learning, using convolutional neural networks, can identify brain tumors from MRI images, offering a promising approach to improving the accuracy and speed of diagnosis. This work presents the identification of brain tumors using convolutional neural networks “CNN”. Thus, our system relies on database augmentation, image pre-processing and extracted features using the CNN criée model and other implemented models such as VGG16,VGG19 and the NasNet exploiting the transfer learning method.