A variety of different video processing modes are supported to accommodate different monitoring and security needs, as described in the table below. Each mode can be customized through the various preferences menus in the SAFR Desktop Client.
| Video Processing Mode | Description | Use Case |
|---|---|---|
| Recognition | Recognize people but don't report events. | Self-Registration kiosk or manned registration terminal. |
| Import | Import faces from video or images. No events are generated. | Batch facial image imports. |
| Learn and Monitor | Learn new faces or recognize already enrolled faces. Events are generated. | High security areas where unknown individuals should be learned, tracked, and differentiated from known persons. |
| Secure Access | Recognize with high confidence and log known persons. Not resistant to photo or video spoofing. | Monitored secure access scenarios such as doors near receptionists. |
| Secure Access with Smile | Recognize and verify liveness using smiles on command. Resistant to photo spoofing. | Non-monitored access for medium security scenario (e.g. internal doors). |
| Secure Access with Liveness | Recognize and verify liveness using motion. Detects changes such as the subject turning their head or walking up to the camera. Resistant to photo spoofing. | Same as "Secure Access with Smile" above but can also be used for walk-through use cases. |
Enrolled Monitoring
| Recognize enrolled faces. Events are generated. | Monitor and alert for persons on watchlists such as VIP lists, known shoplifters, etc. |
| Anonymous Traffic Monitoring | Demographics information collection only. No recognition is performed, and no event images are saved. | Report on demographics trends in retail or other public spaces. |
| Enrolled and Anonymous Traffic Monitoring | Recognize enrolled faces. Demographics information is collected for all people, whether or not they're enrolled. No event images are saved. | Combines the abilities of the "Enrolled Monitoring" and Anonymous Traffic Monitoring" modes above. |
| Enrolled and Unique Traffic Monitoring | Recognize enrolled faces and learn new faces. Events are generated. | Learn faces automatically from video and recognize those faces when seen by cameras to enable use cases such as unique traffic counting or traversal times. |
| Enrolled and Stranger Monitoring | Recognize enrolled faced and generate enrolled person and stranger events. | Medium security areas where unknown individuals should be recorded and available if needed for retrospective analysis. |