Face recognition fails for three groups of reasons: lighting, obstruction, and insufficient face data. All three can be fixed.
Employees — try in this order
- Find a well-lit spot. Standing with your back to a window or light is the number one cause — light from behind turns your face into a dark silhouette.
- Remove your face mask, sunglasses, or a hat covering your forehead.
- Hold the phone at eye level, about an arm's length from your face.
- Look straight into the frame; don't tilt or lower your face.
- Wipe the camera lens — smudges on the lens blur the image.
- Check the Camera permission for the Tanca app in your phone's Settings.
- If it still doesn't work, submit an explanation request and notify your manager.
For AI Cameras at the entrance
- Slow down a little as you pass the camera and look straight ahead.
- Take off your helmet before going through the entrance.
- If the camera doesn't recognize you, don't just keep going — tell your shift manager the same day; don't leave it until the end of the month.
If you have recently changed your appearance significantly — grown a beard, cut your hair short, started wearing new glasses — the system may need you to re-register your face. Notify HR instead of trying over and over.
Administrators — check in this order
| Step | What to check | Where |
|---|---|---|
| 1 | One person or many? Many people at the same entrance points to a device or lighting problem. | Ask the department manager |
| 2 | Has the person had their face registered? | Employee profile |
| 3 | What is the quality of the captured images? | Time & Attendance › Attendance photos (article B6-16) |
| 4 | Is the camera still connected? | Time & Attendance › Settings › AI Camera |
| 5 | Is the Require AI face recognition at shift start toggle turned on? | Time & Attendance › Settings › General |
Fixing it at the installation location
| Symptom | Common cause | Fix |
|---|---|---|
| Dark images, only silhouettes visible | The camera faces a glass door or a strong light source | Change the camera's direction or add a light in front |
| Blurry images | People walk past too fast, or the lens is dirty | Clean the lens and put up a sign asking people to slow down |
| One person is mistaken for another | Not enough face data | Re-register the faces of both people |
| Frequent errors early in the morning or late in the evening | Lighting changes throughout the day | Add a fixed light at the camera location |
During the first week after installing an AI Camera, run it in parallel with the old method and review the Attendance photos screen every day. The misrecognition rate is highest during this period, and it is the cheapest time to adjust the camera position.
