How Our Golden Ratio (φ = 1.618) Face Calculator Works
The FaceScoreLab Golden Ratio Face Calculator runs Google MediaPipe's 478-point facial landmark mesh directly inside your browser, applies automatic eye-line roll correction, and measures five explicit classical proportions against the golden ratio φ ≈ 1.618 (with a clearly disclosed landmark-10 proxy benchmark for upper-forehead height).
Unlike opaque "beauty percentage" calculators that hide how they turn a selfie into a single number, every ratio, exact landmark index, target benchmark, and deviation slope is printed transparently below.
The 5 Golden Ratio Measurements & Exact Formulas
After leveling your eyes horizontally (−rollRad), our browser engine computes these five deterministic ratios:
| Golden Ratio Metric | MediaPipe Landmarks Used | Target Benchmark | Exact Score Formula |
|---|---|---|---|
| 1. Mouth Width / Nose Width | Lip corners (61–291) ÷ alar nose width (129–358) |
1.618 |
`clamp(100 − |
| 2. Eye-to-Chin / Eye-to-Mouth | Vertical eye line to chin (152) ÷ eye line to lip junction (13) |
1.618 |
`clamp(100 − |
| 3. Face Width / Central Eye Span | Bizygomatic width (234–454) ÷ outer-canthus span (33–263) |
1.618 (1 / 0.618) |
`clamp(100 − |
| 4. Lower Third / Lip-to-Chin | Subnasale-to-chin (2–152) ÷ lip-junction-to-chin (13–152) |
1.618 |
`clamp(100 − |
| 5. Face Height Proxy / Cheek Width | Upper-forehead proxy (10–152) ÷ bizygomatic width (234–454) |
1.38* (see note) |
`clamp(100 − |
phiOverall = (Score_1 + Score_2 + Score_3 + Score_4 + Score_5) / 5
Honest Hairline Proxy Note (
1.38*vs1.618): Standard 478-point landmark models stop at the upper frontal forehead boundary (pt[10]) rather than detecting hair roots (trichion). Because the forehead skin landmark sits roughly14%–17%below a typical hairline, an anatomical trichion-to-chin aspect ratio of1.618maps to approximately1.38between MediaPipept[10]and chinpt[152]. We benchmark metric #5 directly against1.38instead of pretending to detect hairlines.
Worked Golden Ratio Calculation Example (Demo Profile)
Click Try Sample Face above to run our built-in local calibration portrait with zero uploads, or verify the math step-by-step on this illustrative profile:
| Ratio Measured | Illustrative Value | Target | Formula Substitution | Sub-Score |
|---|---|---|---|---|
| Mouth Width / Nose Width | 1.54 |
1.618 |
`100 − | 1.54 − 1.618 |
| Eye-to-Chin / Eye-to-Mouth | 1.65 |
1.618 |
`100 − | 1.65 − 1.618 |
| Face Width / Central Eye Span | 1.65 |
1.618 |
`100 − | 1.65 − 1.618 |
| Lower Third / Lip-to-Chin | 1.47 |
1.618 |
`100 − | 1.47 − 1.618 |
| Face Height Proxy / Cheek Width | 1.40 |
1.38 |
`100 − | 1.40 − 1.38 |
| Overall Phi (φ) Fit | — | — | Mean of all 5 ratio scores | 96.2% |
Does a Higher Golden Ratio Score Mean "More Attractive"?
No single number captures human beauty. Classic anthropometric studies show that while faces with balanced proportions often cluster near neoclassical benchmarks (1.50–1.68 across mouth/nose and lower-face ratios), wide, attractive natural diversity exists across sexes, ages, and ethnic backgrounds. Treat your Golden Ratio Fit as an educational framing and styling reference—for example, knowing whether your eye-to-chin proportion is compact or elongated directly informs glasses lens height, fringe depth, and beard/contour framing.
Pair your golden ratio readout with our Facial Symmetry Test (18 bilateral pairs + local Left-Left / Right-Right mirror composites), run the complete 4-pillar Attractiveness Test, inspect fWHR and canthal tilt on the PSL Scale Test, or browse the 42-Metric Sample Full Report.
Verifiable Zero Uploads: All five φ calculations and PNG share-card exports execute locally in your browser. Open DevTools → Network → Fetch/XHR before uploading to confirm zero image or score transmissions.
Community Notes & Questions
0 notes