Multi-Classification of brain tumor based on deep CNN

Abstract

In the last decades, brain tumors are considered one of the mortal cancers in the world. The right tumors detection and identification in the early phases have a significant role to select an accurate treatment. Due to the increasing number of patients and brain tumor types, the manual analyses of Magnetic Resonance Imaging (MRI) images represent a tiring routine and can lead to human errors. In the goal to surpass these problems, an automatic CAD system is needed. We discussed, in this paper, a new model to classify brain tumors using CNN. The suggested scheme is experimentally evaluated on a public dataset. The proposed approach yields a convincing performance compared to previous techniques based on the experimental results.