Main Article Content
A brain tumor is formed by a group of tissue where anomalous cells aregradually added and the most challenging task is classifying brain tumor using magnetic resonance imaging (MRI) to provide treatment for affected patients. Generally, tumor detection and classificationfor brain MRI images are investigated by human. Images are interpreted based on the classification for which several approaches are designed. Information about anatomical patterns and abnormal tissues are obtained from segmentation of brain tumor MRI using the proposedRFCNN segmentation technique. Here the dataset of patients with earlier symptoms of brain tumor has been taken along with their pre-historic medical data. Segmentation has been done for predict the presence of brain tumor whether it is mutation of primary tumor cells (non-cancerous) or secondary tumor cells (cancerous).The proposed neural model can analyse MRI images for detecting the cell mutation and pre-processing of input images to eliminate the parts like skull or vertebral column in advance. The efficiency of the method proposed using a MRI image dataset is compared against existing machine learning and deep learning models. From the results, it is observed that this proposed model obtained a remarkable accuracy, AUC, precision, recall and F-1 score for tumor classification than the methods which used the same dataset.