How accurate is gender recognition?
Gender recognition is 95% accurate on average across male and female categories, measured against ground-truth audits at deployment.
Gender recognition is 95% accurate on average across male and female categories, measured against ground-truth audits at deployment.
How accuracy is measured
For each store, our team samples anonymous face-blurred snapshots from the cameras during the first week of operation and manually verifies the AI’s gender estimate against a human label. The resulting accuracy figure is what we report and what your dashboard reflects.
What the 95% number includes
- It’s an aggregate across all stores, all camera angles, and both gender categories.
- Per-store accuracy can vary slightly based on camera placement, lighting and demographic mix.
- The figure refers to estimation accuracy — Aura Vision does not identify individuals, only counts an anonymous estimated gender per visit.
When accuracy is lower than the average
Stores with steep camera angles, very low light, or short-dwell entrance views can see lower numbers. In those cases we’ll work with you during setup to reposition the camera view or use an alternative angle.