The Application of Using Convolutional Neural Network to Classify MRI Brain Tumor

Brain tumor is a severe disease that requires accurate classification before giving treatment. As traditional diagnosing is time consuming and has a great reliance on experience, the deep learning method is recommended for classifying brain tumor. Deep learning is a newly developed technology that is commonly applied in image recognition field. Many successful applications of deep learning surely increased the efficiency and accuracy of the classification procedure. Deep learning requires sufficient datasets to train the computer, but for brain tumor usually we don't have enough data to input. This paper mainly discussed two methods-data augmentation, and transfer learning to deal with this issue. This paper focused on introducing the types and the effect of data augmentation, then presenting different kinds of transfer learning skills which are applied in different circumstances. Results shows that both transfer learning and data augmentation are advantageous for training the deep convolutional neural network model