Deep Neural Network for Multiple Classification of Brain Tumor Images

Abstract

Classifying brain tumours is important because it helps doctors assess the lesions and choose the best course of therapy. Imaging can be done in a variety of ways to look for brain lesions. However, MRI is widely used because it produces high-quality images and uses no harmful radiation. As a branch of machine learning, deep learning (DL) has lately demonstrated impressive effectiveness, particularly in the areas of classification and segmentation. In this article, we use two freely accessible datasets to suggest a deep learning (DL) model built on a convolutional neural network for the classification of brain tumours. The former divides malignant growths into (meningioma, glioma, and pituitary tumour). The other one classifies gliomas into three distinct stages (Grade II, Grade III, and Grade IV). The first dataset has 3064 T1-weighted contrast-enhanced pictures from 233 patients, while the second dataset has 516 images from 73 patients. The suggested network topology outperforms the state-of-the-art methods in both experiments, with an average accuracy of 96.13% and 98.7%, respectively. The outcomes show that the algorithm can be used to classify various types of brain tumours.