What the term actually describes
A facial symmetry test takes landmark positions, splits the face down the midline, and compares each point against its mirrored counterpart on the other side:
- Eye corners
- Brow height
- Jaw edge
- Mouth corners
Small differences are normal; every real face has some. The score is just a summary of how much left and right diverge in that one frame, not a judgment about which side is "correct."
Why a single photo is the wrong tool for this specific measurement
A mirror comparison only means something if the face was actually square to the camera. Two things distort it before anything about your face changes:
- Head angle — turning a few degrees off-axis (yaw) or tilting slightly (roll) moves one side of the face closer to the lens while the other moves away, changing apparent landmark spacing on each side independently.
- Off-center lighting — a shadow falling across one cheek reads to a landmark model as structure, not illumination.
Of the six ratios Mogalytics tracks, frame symmetry is one of the more pose-sensitive — see the full breakdown of why scores swing between scans for the other camera variables that move a reading the same way. A single uploaded photo can't correct for any of this after the fact.
What a repeatable read looks like instead
The fix isn't a better single photo, it's ruling out head angle and lighting as the variable before trusting the number:
- Face the lens straight-on.
- Keep light even on both sides of the face.
- Take more than one reading — the practice guide covers framing and lighting in more detail.
If your symmetry reading still moves once pose and light are controlled for, that's a more meaningful signal than any single photo can give you.
Where this fits in Mogalytics' composite
Mogalytics tracks frame symmetry as one of six weighted geometry ratios, settled over a 30-frame trimmed-mean window so a single head turn doesn't define the result, alongside:
- Eye line
- Jaw ratio
- Midface
- Cheek ratio
- Bridge
It's read live from your own camera on-device, not from one uploaded photo, so you can check your pose against the lens, adjust, and check again in the same session instead of guessing from a single shot. See what the full composite tier actually means once you have a stable reading, and how to track it across sessions rather than trusting any one scan.
Related reads
Frame symmetry is one piece of a six-measurement composite. If you're auditing the rest:
- Canthal tilt — distorted by the same head-angle problem from a different axis.
- Midface ratio — thrown off by camera distance instead of pose.
- The nose bridge ratio — skewed by angle for a structural reason specific to the bridge's shape.
- Cheekbone ratio shares this test's head-turn sensitivity, while jaw ratio is pitch-sensitive instead.
- Golden ratio face — the proportion template the critique literature says symmetry and averageness outpredict.