What a golden ratio face test measures
The golden ratio, written φ (phi), is about 1.618. Two lengths are in this ratio when the longer divided by the shorter equals their combined length divided by the longer. A golden ratio face test compares ratios of facial measurements with this number.
This tool uses a single front-facing photo. It finds 478 facial landmarks with Google’s MediaPipe Face Landmarker, levels your eye line, and calculates 8 ratios: three width ratios, four height ratios and one lip ratio. You get:
- A golden ratio score from 0 to 100: the average of the eight ratio scores, each counted equally.
- Each ratio’s value, its percentage difference from φ and its individual score.
- The ratio closest to φ and the one furthest from it.
- The measurement lines on your photo: gold lines for the widths and blue lines for the heights.
These measurements come from a single 2D photo. Head angle, expression and camera distance can affect them.
The eight ratios
The width and height ratios are based on the golden-proportion method that orthodontist Robert Ricketts proposed in 1982, as described by Mantelakis and colleagues (2018). We use the ratios our landmark model can measure in a front-facing photo.
Widths
Widths are measured between pairs of landmarks:
- Eye span ÷ mouth: the distance between your outer eye corners, divided by the width of your mouth from corner to corner.
- Mouth ÷ eye gap: your mouth width divided by the distance between your inner eye corners.
- Mouth ÷ nose: your mouth width divided by the width of your nose, measured across the outer edges of your nostrils.
- Outer eye span ÷ mouth width
- 1.64 · 1.6% above φ
- Mouth width ÷ inner eye gap
- 1.49 · 7.8% below φ
- Mouth width ÷ nose width
- 1.21 · 25.5% below φ
Heights
After leveling your eye line, the tool measures vertical distances between four reference levels:
- Eye line: the level of the midpoint between your outer eye corners.
- Nose base: the average level of the outer edges of your nostrils.
- Mouth: the line where your lips meet.
- Chin: the lowest point of your chin on the midline.
The four height ratios are:
- Eye line–chin ÷ eye line–mouth
- Eye line–chin ÷ nose base–chin
- Nose base–chin ÷ mouth–chin
- Eye line–nose base ÷ nose base–mouth
- Eye line to chin ÷ eye line to mouth
- 1.67 · 3.3% above φ
- Eye line to chin ÷ nose base to chin
- 1.47 · 9.0% below φ
- Nose base to chin ÷ mouth to chin
- 1.69 · 4.5% above φ
- Eye line to nose base ÷ nose base to mouth
- 1.16 · 28.5% below φ
Lips
The last ratio divides the height of your lower lip by the height of your upper lip, each measured at the center. We include this comparison because the golden ratio is often discussed in relation to lip proportions; it is separate from Ricketts’ original ratios.
Research has explored other lip proportions, too. In one experiment with altered photos, raters preferred an upper-to-lower lip ratio of 1:2 among the options tested. That study compared lip surface areas, so its result is not directly interchangeable with the lip heights measured here.
How the golden ratio score works
Each ratio is scored by how close it is to φ, with equal treatment of differences above and below it. A ratio equal to φ scores 100. A ratio 20% above or below φ scores about 82. Small differences have little effect on the score:
| Difference from φ | Ratio | Score |
|---|---|---|
| 0% | 1.618 | 100 |
| 5% | 1.54 or 1.70 | 99 |
| 10% | 1.46 or 1.78 | 95 |
| 20% | 1.29 or 1.94 | 82 |
| 30% | 1.13 or 2.10 | 64 |
| 40% | 0.97 or 2.27 | 46 |
Your golden ratio score is the average of the eight ratio scores, rounded to a whole number. Each ratio contributes equally to the final score.
We chose how quickly scores fall as ratios move away from φ. This scale has not been calibrated against population data. The AI attractiveness test uses the same golden ratio score as one of its five components, with a weight of 10%, so both tools give the same score for the same photo.
Resizing a photo without changing its proportions leaves these ratios unchanged. Camera distance and head angle can still affect them: close-up photos distort perspective, while tilting or turning your head changes the apparent distances between features.
Reading your results
The score combines the eight individual ratio scores. Below it, you’ll see the average percentage difference from φ, calculated without canceling out differences above and below it. Photos with the same overall score can have different individual ratios, so check the breakdown as well.
Closest to φ and furthest from φ identify the ratios with the smallest and largest percentage differences from 1.618.
Each ratio row shows:
- The measured value, to three decimal places.
- How far it is above or below φ, as a percentage of φ. Differences under 1% are labeled “Near φ”.
- That ratio’s score.
- A bar showing the relative size of the two measurements, with the first in blue. The gold mark shows the split for a 1.618:1 ratio: 61.8% to 38.2%. If the blue section extends past the mark, the first measurement is more than 1.618 times the second.
Above or below φ describes the ratio’s value. For example, a mouth ÷ nose ratio below φ means your mouth is less than 1.618 times the width of your nose. Neither direction indicates a better-looking feature.
Why face length to width isn’t included
Some golden ratio face calculators start with face length divided by face width, measuring length from the hairline to the chin. Our landmark model does not identify the hairline. Its uppermost point on the facial midline is on the forehead, so using it would underestimate that length. We therefore leave this ratio out.
The tool also leaves out Ricketts’ ratios that depend on temple edges or nasal bridge boundaries the model does not reliably identify. Our face shape detector does measure length to width using the forehead point. It compares that measurement with its face-shape profiles.
Does the golden ratio make a face attractive?
Research has not established 1.618 as a universal standard for facial beauty. Studies have reached different conclusions:
- The original claim. Ricketts (1982) reported golden proportions in faces he considered ideal, including photos of models. Later studies tested whether closeness to these ratios relates to how attractive a face is rated.
- Some support. In a study of magazine cover models and former orthodontic patients, Pancherz and colleagues found wide variation from the predicted golden-proportion measurements. Among the former patients, those rated attractive had smaller deviations than those rated unattractive.
- Limited agreement in models. Mantelakis and colleagues (2018) measured twelve facial ratios in photos of 40 Black professional models. Only one ratio did not differ significantly from the golden proportion.
- Other proportions. In Pallett, Link and Lee’s experiments using female faces, ratings were highest when the eye-to-mouth distance was about 36% of face length and the distance between the pupils about 46% of face width. These findings concern different measurements from the ratios used in this tool.
- Review findings. A 2024 review by Naini concluded that there is no convincing evidence linking the golden ratio to idealized human proportions or facial beauty, and none supporting its use in facial surgical planning.
Marquardt’s golden-ratio mask has been criticized by Holland (2008) for reflecting masculine facial proportions seen in European fashion models and fitting other populations poorly. This tool measures individual ratios rather than matching your face to that template.
A higher score here means your measured proportions are closer to 1.618 in this photo. It does not measure your attractiveness or suggest that a feature needs changing.
Where the golden ratio comes from
The ratio was defined in geometry long before anyone applied it to faces. Around 300 BC, Euclid described dividing a line in “extreme and mean ratio”, so that the whole line is to the longer part as the longer part is to the shorter. That proportion is φ. In 1498, Luca Pacioli called it the divine proportion in his book De Divina Proportione.
Many popular claims about the ratio do not hold up. The Naini review notes that Leonardo da Vinci did not use the golden ratio in his drawings of human proportions, and that Pacioli did not present it as the secret of beauty. It also questions the familiar diagrams that fit golden rectangles onto buildings such as the Parthenon.
Limitations
- A single 2D photo. The ratios cannot capture facial depth, and perspective affects the apparent proportions.
- Landmarks are estimates. The model estimates where each point is. Shadows, facial hair, glasses or hair over the face can shift them, and the height of thin lips is especially sensitive to small errors.
- Some reference points are approximations. We use the average height of the outer nostril edges for the nose base and the line where the lips meet for the mouth. These may differ from reference points identified by a clinician.
- One reference for everyone. The same number, φ, is used for every face regardless of age, sex or ancestry, although typical facial proportions vary between populations.
- Not a medical or surgical assessment. The results should not be used to plan treatment.
How to get a fair measurement
For more consistent results across photos:
- Face the camera straight on, with your head level: don’t tilt your chin up or down. Vertical tilt changes the height ratios.
- Avoid close-up selfies. Hold the camera at least an arm’s length away, or have someone else take the photo. Close-range photos can distort facial proportions, especially the nose.
- Keep a neutral expression with your mouth closed. A smile widens your mouth and thins your lips, and seven of the eight ratios use your mouth or lips.
- Use even light from in front of you. Strong shadows can make it harder to locate the landmarks around your nose and lips.
- Pull hair off your face and remove glasses, so the eye corners and jawline are clear.
If two photos give different results, check head angle, camera distance and expression first. A change in score alone does not show that your face has changed.
Sources
- Ricketts, R. M. (1982). The biologic significance of the divine proportion and Fibonacci series. American Journal of Orthodontics, 81(5), 351–370.
- Mantelakis, A., Iosifidis, M., Al-Bitar, Z. B., Antoniadis, V., Wertheim, D., Garagiola, U., & Naini, F. B. (2018). Proportions of the aesthetic African-Caribbean face: Idealized ratios, comparison with the golden proportion and perceptions of attractiveness. Maxillofacial Plastic and Reconstructive Surgery, 40, 20.
- Pancherz, H., Knapp, V., Erbe, C., & Heiss, A. M. (2010). Divine proportions in attractive and nonattractive faces. World Journal of Orthodontics, 11(1), 27–36.
- Pallett, P. M., Link, S., & Lee, K. (2010). New “golden” ratios for facial beauty. Vision Research, 50(2), 149–154.
- Holland, E. (2008). Marquardt’s Phi mask: Pitfalls of relying on fashion models and the golden ratio to describe a beautiful face. Aesthetic Plastic Surgery, 32(2), 200–208.
- Naini, F. B. (2024). The golden ratio—dispelling the myth. Maxillofacial Plastic and Reconstructive Surgery, 46, 2.
- Popenko, N. A., Tripathi, P. B., Devcic, Z., Karimi, K., Osann, K., & Wong, B. J. F. (2017). A quantitative approach to determining the ideal female lip aesthetic and its effect on facial attractiveness. JAMA Facial Plastic Surgery, 19(4), 261–267.
- Ward, B., Ward, M., Fried, O., & Paskhover, B. (2018). Nasal distortion in short-distance photographs: The selfie effect. JAMA Facial Plastic Surgery. doi:10.1001/jamafacial.2018.0009.
- Kartynnik, Y., Ablavatski, A., Grishchenko, I., & Grundmann, M. (2019). Real-time facial surface geometry from monocular video on mobile GPUs. arXiv:1907.06724.
- Google AI Edge. Face landmark detection guide. Documentation for the 478-landmark model used by this tool.
