API Connected TF 2.10 ONNX FP32 Grad-CAM XAI
MRI Input
⬆
Drop MRI image here
PNG, JPG, JPEG — any brain MRI scan
MRI preview
Inference Backend
(TF backend only)
Quantization Info
Dynamic INT8
Static INT8
Compare
Dynamic Quantization converts model weights to INT8 at export time. Activations remain FP32 at runtime and are quantized on-the-fly. No calibration data needed.
Static Quantization quantizes both weights and activations to INT8. Requires calibration data. It can be smaller and faster, but accuracy may vary by architecture.
Loading benchmark...
Analysis Results
🔬
Upload an MRI scan and click analyse
Supports glioma · meningioma · no tumor · pituitary
Class Probabilities
—
Confidence
—
Latency
—
Backend
Class Reference
ClassDescriptionSeverity
Glioma Tumor of glial cells High
Meningioma Tumor of meninges Medium
No Tumor Normal brain MRI None
Pituitary Pituitary gland tumor Medium
Model Performance Comparison
Loading model info...