Brain tumor diagnosis from MRI scans is time-sensitive and expertise-dependent. Early and accurate classification can directly impact treatment decisions and patient outcomes. This project builds an end-to-end deep learning pipeline for automated brain tumor classification.
EfficientNetB0 was selected for its optimal accuracy-efficiency tradeoff. The pipeline covers data preprocessing, augmentation, transfer learning from ImageNet weights, fine-tuning, and evaluation. Built end-to-end with a focus on reproducibility.
92% accuracy on the test set. The model demonstrates strong performance across tumor classes, with evaluation metrics including precision, recall, and F1 score per class.
“Accuracy alone is not enough in medical AI — but it is where you start.”