Multi Classification of Brain Tumor Detection Using MRI Images: Deep Learning Approach

Abstract

Brain tumor detection at early stages is very important for successful treatment. The life of the patient can be saved if it is detected in the early stages and also useful for proper and efficient medication. Nowadays, machine learning algorithms are being used for diagnosing purpose in the medical field. Computers always give better result compared to manual diagnostic. Brain tumor detection and segmentation is a very crucial task, as manual processing of medical images leads to the wrong prediction. Magnetic Resonance Imaging scans have proved to be helpful for the diagnosis or segmentation of brain tumors. The image segmentation process is used for the extraction of tumors that are not normal in the brain. With the use of efficient data mining techniques and different classification algorithms, prediction of the disease can be performed at an early stage with better accuracy and effectiveness. In the field of medicine, Machine Learning and Data Mining techniques have proved to be effective and useful for better prediction. Deep learning is one of the subparts of machine learning and recently proved to be of high importance in classification and segmentation-related problems. This work is implemented using a deep learning approach, modeled into Convolutional Neural Network to classify the results into different categories like “Meningioma”, “Glioma”, “Pituitary” or “TUMOR NOT FOUND”.