suspicious 0.86🕵️
tracking🎯
flagged🚩
Suspicious behaviour scored against the scene's normal baseline.
What it does
An anomaly-detection system that learns normal behaviour in a scene and flags suspicious actions associated with theft.
How it works
- Step 1. The model learns the normal motion and interaction patterns for a space.
- Step 2. Each frame is scored against that baseline.
- Step 3. Suspicious sequences (concealment, grab-and-go) raise the anomaly score.
- Step 4. Crossing the threshold triggers a flagged clip for review.
Where it's used
- Retail loss prevention
- Self-checkout monitoring
- Warehouse security
- Parking lots
Want something like Theft Detection?
Tell us your use case — we'll scope a build that fits.