A new deep CNN for brain tumor classification

Abstract

In the last years, the brain tumor is considered as one of the most deadly tumors around the world. It can affect adults and children. The wrong classification of the tumor brain will lead to bad consequences. Consequently, the right identification of the type and the grade of tumors in the early stages has a significant role to choose a precise treatment plan. Due to the various brain tumor types and the big amounts of data, the manual technique for examining the Magnetic Resonance Imaging (MRI) images becomes time-consuming and can lead to human errors. Therefore, an automated Computer Assisted Diagnosis (CAD) system is needed to overcome these problems. We suggested a new CNN scheme to classify different brain tumors. The suggested model is experimentally evaluated on a benchmark dataset. Experimental results affirm that the suggested approach provides convincing results compared to existing methods.