Hybrid CNN-Transformer Based Brain Tumor Detection and Classification Using MRI: A Novel Framework with Multi-Level Attention and Regional Explainability
A novel hybrid convolutional neural networks and transformer architecture, HybCT-Net, augmented with a multi-level attention module and a regional explainability pipeline for brain tumor detection and classification is proposed, demonstrating superior performance than contemporary CNN, transformer and hybrid baselines.