Design of automated computer-aided classification of brain tumor using deep learning

Abstract

In the recent years, health issues have inescapably become center of attention for many researchers. Brain tumor is now a leading cause of death among medically certified deaths. Brain image diagnosis is manually examined by the neurologist. It is time consuming and may lead to errors. The general idea of this research is to analyze the brain tumor based on magnetic resonance imaging (MRI) of medical images. The design of this system is aimed at detecting the brain tumor classifying the MRI samples. The system uses computer-based procedures to detect tumor blocks and classify the type of tumor to normal, benign, and malignant using tensor flow in MRI images of different patients. A promising method to perform the design is through a deep learning process. Deep learning is currently a well-known and superior method in the pattern recognition field. The performance measure for detection would be Equal Error Rate (EER), false acceptance rate (FAR), and false rejection rate (FRR). The higher percentage of accuracy of the biometric system depended on how much lower the ERR value would be. The samples are already available from a standard database, Multimodal Brain Tumor Image Segmentation Benchmark (BraTS). A comparison had been done between two different methods for classification of the brain tumor.