Enhanced Brain Tumor Classification Using a Modified Xception Convolutional Neural Network

Brain tumors are abnormal cells in the brain, and some types can lead to cancer. This study introduces a deep learning method that uses neural network (CNN) to analyze brain tumor images. We utilize transfer learning with the Xception model, which has been trained on the Image Net dataset, to effectively manage the complexities of medical imaging data. Our methodology includes image preprocessing, feature augmentation, and normalization techniques to improve image quality. We evaluated several models, including MobileNetV2, VGG16, Xception, and VGG19, and achieved the highest accuracy of 96.15% with the Xception model.