Brain Tumor Classification Deep Learning Model Using Neural Networks

The timely diagnosis of brain tumors is currently a complicated task. objective was to build an image classification model detect the existence or not by adding header ResNet-50 architecture. CRISP-DM methodology used for data mining. A dataset 3847 MRI images used, 2770 training, 500 validation, and 577 testing. were resized 256 × scale then generator created that responsible dividing pixels 255. training performed evaluation process carried out, obtaining accuracy percentage 92% precision 94% in process. It concluded proposed CNN composed head with ResNet50 architecture seven-layer convolutional network achieves adequate accuracy, becoming efficient complementary proposal other models developed previous works.