Machine Vision & Industrial Inspection AI worked example

AI Inspection Precision Recall at 71% required detection recall specification: a worked example

Suppose required detection recall specification falls to 71%. This page works the full calculation at that level so you can see exactly which result moves and by how much. Calculate the recall (sensitivity) of an AI inspection model from a validation sample by dividing true positives by the total number of actual defects, and see how far the result is from your required detection specification.

The inputs for this scenario

  • True positives (defects correctly detected): 96 defects (held at the documented default)
  • Total actual defects in validation sample: 100 defects (held at the documented default)
  • Required detection recall specification: 71 % (the input this scenario stresses; the baseline uses 98)

Working through the calculation

  • The calculation starts from the formula this tool documents: Recall = true positives / total actual defects x 100.
  • AI model recall (detection rate) works out to 96 % at these inputs, and this is the headline figure for the scenario.
  • Gap to required recall specification works out to -25 points at these inputs.
  • True positives detected works out to 96 count at these inputs.
  • Total actual defects in sample works out to 100 count at these inputs.

How this compares with the baseline

  • Against the tool's baseline example, where required detection recall specification sits at 98% and the headline result is 96 %, this scenario lands almost exactly on the baseline at 96 %.
  • It computes detection recall as true positives over total actual defects, and the gap in percentage points to your required recall specification. When the numbers land here, the stressed input is the lever to work; the walkthrough above shows exactly how much each output recovers as it climbs back toward the baseline.

Results at a glance

  • AI model recall (detection rate): 96 % (headline result)
  • Gap to required recall specification: -25 points
  • True positives detected: 96 count
  • Total actual defects in sample: 100 count

Run it with your numbers

  • To rerun this with your own numbers, open the live AI Inspection Precision Recall calculator, set required detection recall specification to your actual value, and adjust the remaining inputs to match your operation.

Last reviewed 2026-05-12.