Take two photos. The app estimates your age and gender, shows how three datasets classify you, and explains what changed between your photos.
Look straight at the camera in good, even light. Remove glasses and hats.
Now change one thing: the light, your angle, or your distance from the camera.
Make sure your whole face is visible in both.
The app will show how each dataset classifies you: FairFace, UTKFace, LFW and this project's own balanced model. It also compares two recognition models (a general one and our own) on your photos, so you can see where they agree or disagree.
1:00
Still working — the first analysis of a session can take up to a minute while the models warm up.
Pick a dataset to see how a model trained on it classifies your first photo.
Two recognition models, same two photos. Do they agree?
Appearance predictions from the LFW/LFWA+ attribute model.
These describe differences detected in the images, not permanent characteristics of the person. Lighting, pose and image quality can affect these predictions.
| Measure | Photo 1 | Photo 2 |
|---|
Your two photos and results have been cleared from this device.