A Review on Brain Tumour Detection and Classification Using Machine Learning

Abstract

A brain tumor is a malformed tissue in which cells multiply rapidly and uncontrollably, and the growth of the tumor cannot be stopped. In this specific study, a multimodal convolutional neural network (CNN) is developed to do classification, and the deep neural network (DNN) method is applied for the automated identification of brain cancers using magnetic resonance imaging (MRI) images of the human brain. Brain tumor classification and detection modules provide all diagnostic data. Four datasets provided accuracy and loss results for validation and training. The suggested technique for diagnosing brain tumors relied on deep learning (DL) and made use of brain magnetic resonance for tumor assessment using automated categorization methods. In this study, a multimodal CNN model is employed to enhance images and enhance classification. On four datasets, while employing multimodal CNN for classification and DNN for detection. Data Set 1 achieves a training accuracy of 98.18 percent and a validation accuracy of 98 percent. The second dataset yields 98% validation accuracy and 99.88% training accuracy. The third dataset has a validation accuracy of 99.47% and a training accuracy of 99.88%. The final dataset has a validation accuracy of 64.91 percent and a training accuracy of 78.97 percent. This method could classify even rare tumor types with extra data.