Brain Tumour Detection Using Resnet 50 and Mobilenet

Abstract

The scientific community defines a brain tumour as  a mass or growth of abnormal cells in the brain. A brain tumour is a development of abnormal cells, some of which may develop  into cancer. MRI scans are the most common way to find brain  tumours and are used to detect brain cancer. There are different  types of tumours exist. They are cancerous(malignant)and non- cancerous(benign) in the brain identification of unchecked tissue growth in MRI may help us diagnose brain cancer. Machine Learning and Deep Learning algorithms are used to identify this tissue growth. When these algorithms are applied to MRI scans, a faster prediction of brain tumours is made, and   a better degree of accuracy aids in treating patients. MRI scans   allow us to perform rapid analysis and identify the exact location of unwanted tissue growth. Various uses include image   recognition and identifying objects, image classification, segmentation, neural network and data processing. The proposed model successfully classified the MRI image into four   classes: glioma, meningioma, and pituitary tumour and no tumour, indicating that the given brain MRI has no tumour. In  this paper the proposed models are MobileNet and Resnet50 and gives accuracy of 0.98. These models classifies the type of tumour very accurately.