Brain Tumor Segmentation/Detection using Transfer Learning with VGG19

Using transfer learning, we can reuse the previously trained model. Transfer learning involves utilizing knowledge added from an earlier activity. It is most widely used in the three domains i.e. image prediction, image classification, and natural language processing. Nevertheless, training deep learning models from scratch on huge datasets may lead to high computational complexity. To solve this issue, this paper uses transfer learning for image segmentation and prediction. A pre-trained model of VGG19 is employed in this study. The VGG19 is utilized for image segmentation and prediction. The model is implemented on the Google Colab platform because the performance of the model rests on the GPU of the system. The result using the VGG19 model for the task at hand achieved an accuracy of 98.29%.