How it works
What the editor is actually doing
When you upload a photo, two models run locally. A face landmark model locates your chin, your eye line and the horizontal centre of your face. A segmentation model separates you from the background, which we then use to find the true top of your head, hair included. That last part matters: landmarks alone can only find your hairline, and hair volume varies enough to push a photo out of tolerance on its own.
With chin and crown located, the crop becomes arithmetic. The head is scaled so its chin-to-crown height lands in the middle of the country's permitted range, the eye line is placed at the expected height, and the frame is rendered at the exact pixel dimensions for the format. You get the resulting measurement in millimetres rather than being asked to trust the output.




