Classification of Brain Tumors via Deep Learning Models

Brain tumors threathen human health significantly. Misdiagnosis of these tumors decrease effectiveness of decisions for intervention and patient's state of health. The conventional method to differentiate brain tumors is by the inspection of magnetic resonance images by clinicians. Since there are various types of brain tumors and there are many images that clinicians should examine, this method is both prone to human errors and causes excessive time consumption. In this study, the most common brain tumor types; Glioma, Meningioma and Pituitary are classified using deep learning models. While the main objective of this study is to have a high rate of accuracy, the time spent is also examined. The aim of this study is to ease clinicians work load and have a time efficient classification system. The system which has been built has an accuracy up to 90%.