EfficientNetB3 Brain Tumor MRI Classifier

4-class T1-weighted brain MRI classifier (glioma, meningioma, notumor, pituitary). Reproduces the best model from an IEEE 2024 paper (DOI: 10.1109/ICC-ROBINS60238.2024.10533941) and extends it with production evaluation. Input: 256x256x3 RGB, raw [0,255] pixels (EfficientNet rescales internally). Files: efficientnetb3.onnx (serving), EfficientNetB3_model_weights.h5 (Grad-CAM), temperature.json (calibration T).

Not a medical device — research/education only; dataset has documented train/test leakage. Full evaluation + code: https://github.com/Udit013/Brain_Tumor_Classificatioin

Downloads last month
-
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Space using Udit013/brain-tumor-efficientnetb3 1