Medical Scan Classifier (2D) — CNN Forward Pass

Scan case

Generate a single case

Scan speed: 1.0× Feature map

Batch testing

Lesion prevalence: 50% Severity mix (advanced share): 50% Batch size

Diagnosis threshold

Classify positive above: 0.50

P(lesion) = 00.001

Stats

Cases run0

Sensitivity (TPR)—

Specificity (TNR)—

Accuracy—

AUC—

Threshold0.50

	Pred +	Pred −
Actual +	0	0
Actual −	0	0

A fixed 3×3 blob-detector kernel slides across the synthetic scan, smoothing local intensity. The strongest (max-pooled) response feeds a sigmoid to produce the diagnosis probability shown above — plotted flat here with plain Canvas2D instead of a 3D scene.