An Effective Application to Identify Brain Tumor using Deep Learning Model

Abstract

Brain tumor is one of life threatening diseases for humans and the treatment is challenging. Recently the disease diagnosis industry is seeing enormous developments. Brain tumors can be identified from Magnetic Resonance Imaging (MRI) images. There are existing techniques available for brain tumor detection using image processing techniques. Some recent studies used machine learning approaches for brain tumor detection. However, an effective model and application is required for this life threatening disease. Availability of dataset is an added advantage for these studies. Nowadays, large amounts of data can be preserved for research and these can be used effectively by deep learning models. Disease diagnosis through deep learning techniques are emerging these days. In this paper, brain tumor detection is proposed through a deep learning model, Convolutional Neural Network (CNN). Deep learning models are achieving good results on brain tumor detection. In this work, an application is proposed, in which users can upload the MRI image and detect whether it is a tumor or normal MRI. CNN based classification for brain tumor detection has achieved highest classification accuracy around 99.5%. Experimental results showed that high precision value 99.3% for optimized training epochs.