Brain Tumor Identification Based on AlexNet and VGG

Abstract

Early diagnosis of brain tumor symptoms through Computer Tomography (CT) imaging is an important means of treating brain tumors. However, doctors who identify and diagnose brain tumors with naked eyes may cause misdiagnosis due to overwork or inexperience, which greatly affects the rehabilitation of brain tumor patients. In recent years, there is a lack of a clinical application tool for automatic identification and classification of brain tumor CT images to help doctors improve the rate of brain tumor treatment. This study proposes an approach for identifying and classifying brain tumors using a modified AlexNet and Visual Geometry Group (VGG). The methodology involves the comparative analysis of the performance of the modified AlexNet and VGG in identifying and classifying brain tumors. Our findings reveal that the modified AlexNet demonstrates superior performance, with an accuracy rate of 69.54%, in comparison to the VGG. The proposed approach holds tremendous potential for future clinical application as an auxiliary tool to aid medical professionals in accurately diagnosing brain tumors.