Oral commissure tilt
Angle of the line between the mouth corners relative to the interpupillary line, plus the vertical position of the corners relative to the lip midline.
How to read it
the corners of the mouth sit below the lip midline at rest (downturned commissures). This is strongly age- and expression-dependent.
the mouth corners sit approximately level with the lip midline.
the corners of the mouth sit above the lip midline at rest.
A residual smile of even a few percent (see blendshape values in capture quality) shifts this by several degrees.
Reference values by population
| Reference population | Female (mean ± SD) | Male (mean ± SD) | n | Source |
|---|---|---|---|---|
| North American White | +0° ± 3.5 | +0° ± 3.5 | — | Neutral-expression convention: level commissures ≈ 0°; no population table available prov. |
| East Asian | +0° ± 3.5 | +0° ± 3.5 | — | Neutral-expression convention: level commissures ≈ 0°; no population table available prov. |
| South Asian | +0° ± 3.5 | +0° ± 3.5 | — | Neutral-expression convention: level commissures ≈ 0°; no population table available prov. |
| African / African diaspora | +0° ± 3.5 | +0° ± 3.5 | — | Neutral-expression convention: level commissures ≈ 0°; no population table available prov. |
| Middle Eastern | +0° ± 3.5 | +0° ± 3.5 | — | Neutral-expression convention: level commissures ≈ 0°; no population table available prov. |
| Hispanic / Latino | +0° ± 3.5 | +0° ± 3.5 | — | Neutral-expression convention: level commissures ≈ 0°; no population table available prov. |
| Not specified — show pooled | +0° ± 3.5 | +0° ± 3.5 | — | Neutral-expression convention: level commissures ≈ 0°; no population table available prov. |
Where these numbers come from
Transcribed from the cited anthropometric literature; SDs of ratios are propagated conservatively (independence assumed). Entries marked prov. await line-by-line audit against the primary tables — see the norm provenance note.
Why every value has an error bar
Every value Canthion reports carries a 1σ uncertainty propagated from MediaPipe Face Mesh's published accuracy (3.96% of inter-ocular distance mean absolute error per landmark, Kartynnik 2019). For a typical 300 px inter-pupil photo that is ≈12 px per landmark, which is why small ratios and angles get wide intervals. A number without an interval is a number pretending to be more precise than it is.
References
- Kartynnik, Ablavatski, Grishchenko & Grundmann (2019). Real-time Facial Surface Geometry from Monocular Video on Mobile GPUs. arXiv:1907.06724. arxiv.org/abs/1907.06724
Related: how bone-level measurements are taken from a photo · methodology · all insights