Deep Dive into Brain Tumor Classification: Transfer Learning with Convolutional Neural Networks

It is important to identify brain tumors and their type and stage and choose the most suitable treatment. In therapeutic therapy, brain tumors are identified via several diagnostic tests. This research study focuses on assessing and comparing the CNN pre-trained models' accuracy for brain tumor classification and identification using deep transfer learning approaches and leveraging the transfer learning-based Convolutional Neural Networks (CNNs) pre-trained models, including VGG19, VGG16, InceptionV3, ResNet50, and InceptionResNetV2. We utilized the dataset of brain tumor images from Kaggle. The experimental findings show that the VGG16 pre-trained model has the highest accuracy (90%) compared to other CNN pre-trained models. The findings promise significant contributions to medical image analysis, fostering advancements in AI-powered tools for better brain tumor diagnosis and patient care.