Multicalss Brain Tumor Recognition Using Convolutional Neural Network

Abstract

Day by day the use of computers is increasing everywhere. Computer science is already widely used in the medical field. Brain tumor diagnosis can be performed using computer vision. This research applied several algorithms on this. Among them is the Convolution Neural Network (CNN), which is the most common application of CNN for image recognition challenges. This has been used with MRI images. The outcomes are fairly decent. Also applied VGC-16, MobileNetV2, and InceptionV3 on brain tumor MRI images. The experiment is conducted on a dataset of 3264 images containing four different types of brain tumor (glioma, meningioma, pituitary, and no tumor). It is tough to collect data from hospitals so primarily we collected data from online resources. For experiment first I preprocessed the image and then apply those algorithms. Among them, CNN gives us the best accuracy which is 95%. This study states the various algorithms performed on the same datasets and the best model.