How it works
Customer chooses which facial images to add to a watch-list. The system creates a unique signature representing the features of each face.
Video feeds from cameras are analyzed in real-time. Faces found in the crowd are compared for similarity against those in the watch-list.
Similar facial images above the review threshold are raised as alerts to users. All other facial images are automatically discarded immediately.
Users review the alerts, comparing facial images from cameras against the corresponding watch-list images to determine whether further action is needed.