Deep learning approach for detecting and localizing brain tumor from magnetic resonance imaging images

<span lang="EN-US">Brain is the most important part of nervous system. Brain tumor mainly a mass or growth abnormal tissues in brain. Early detection brain can reduce complex treatment process. Magnetic resonance images (MRI) are used to detect tumor. In this paper, we have introduced deep convolutional neural network (CNN) automatic segmentation using MRI medical which solve vanishing gradient problem. Classifying with Resnet-50 and InceptionV3 order identify whether there not. After step, compared accuracy level both CNN models. Thereafter, applied U-Net architecture individually encoder avieved promising results. The publicly available low grade gliomas (LGG) dataset has been utilized test model. Before applying model on preprocessing several augmentation techniques done obtain quality dataset. U-net provided 99.55% accuracy. On other hand, our proposed method ResNet-50 showed 99.77% accuracy.</span>