Project Overview:
This project explores the use of different deep learning architectures to improve the classification of brain tumors from medical images, with a focus on efficiency, accuracy, and generalization. We evaluate traditional convolutional neural networks (CNNs), residual networks (ResNets), Topological Data Analysis (TDA) CNNs, and Vision Transformers (ViT), comparing their performance on a publicly available MRI dataset. The goal is to identify the model that offers the best balance of computational efficiency and classification accuracy, with potential applications in clinical settings where quick and reliable tumor detection is critical.