A Comparative Study of Deep Learning Models for Brain Tumor Classification

Brain tumor is a fetal disease nowadays that leads to cancer, so detecting brain tumors at an early stage is very crucial. There are many deep neural networks available in the field of medical image processing to detect brain tumors. This paper compares four deep-learning techniques for classifying brain tumor. The models that are used in this model are DenseNet-121, ResNet-50, VGG-16 and Inception-V3. The performance evaluation is based on accuracy, precision, F1-score, AUC ROC score, and Cohen Kappa score on the brain-tumor-detection-mri Dataset from Kaggle. A total of 2400 + images belonging to two classes. We have an accuracy of $\mathbf{9 4. 5 \%}$ for DenseNet-121, 97.5% for ResNet-50, and 94% for VGG-16 and Inception-V3, and we have a score of $73 \%$. The results provide insights into the strengths and weaknesses of each technique and offer guidance for selecting the most suitable approach for image classification in the medical imaging domain.