MRI-Based Brain Tumour Detection Utilizing Machine Learning and Deep Learning Techniques

Advance identification of brain tumors is necessary. Brain tumors can be categorized by biology and can be treated by conclusive brain surgery only. Computer-aided approaches can assist doctors to categorize brain tumors. Here, some machine and deep learning techniques are accustomed to detecting three types of tumors and also including healthy brain, using boom imaging. brain MRIs so doctors can diagnose tumors. and very accurate in the first stages. Here we consider a collection of data containing 3264 MRI scans with different types of tumors was included in research. Primarily, pre-processing and optimization algorithms used on MR scans of the brain. Then, build a 2-D convolutional neural network (CNN) and autoencoder network, which were pre-trained by our established Meta Algorithms. So, a 2D CNN has many convolution layers. Every layer from hierarchical network is $\mathbf{2 x 2}$ kernel functions. The network contains of four merging and eight rotating layers, and after all the rotating layers, buffer layers are used. The advance autoencoder network is a convolutional autoencoder network and a convolutional network for categorization, which utilize last layer of the output encoder of the primary stage. In addition, $\mathbf{6}$ machine learning methods used to clarify brain tumors were distinguish in this research.