Free Symmetrical Face Test — Score Your Facial Symmetry

How the Symmetrical Face Test Works

1. 01 · Choose a photo

   Use even lighting, a relaxed expression, and a camera at eye level.

2. 02 · Map the landmarks

   MediaPipe detects 478 landmarks locally. Six pairs use the homepage’s nose-centered reference line.

3. 03 · Explore the result

   See your score, mirrored comparisons, and largest differences. First-run model loading depends on your device and connection.

What Facial Symmetry Actually Measures

This facial symmetry test measures differences in a photograph. Its score compares horizontal distances from paired landmarks to a vertical line through the nose landmark. It does not reconstruct your face in three dimensions.

Pair differences are expressed as percentages of landmark-based face width. Larger values mean greater horizontal mismatch in that photo. There is no validated threshold here separating a normal difference from a concerning one.

The Symmetry Score Formula

This page uses the exact symmetry dimension in the FaceScored scoring formula. Each of the six pairs has equal weight: one sixth. There are no separate eyebrow or nose weights.

Let c be the nose landmark’s horizontal position, xL and xR the paired positions, and W the distance between face-width landmarks 234 and 454.

Pair deviation = |(c − xL) − (xR − c)| / W
A = sum of six pair deviations / 6
S = round(clamp(98 − 520 × A, 0, 100))

The pairs are 33/263, 133/362, 61/291, 234/454, 93/323, and 172/397. Distances use normalized two-dimensional landmark coordinates. The reference constant is 98 and the deviation multiplier is 520; the maximum produced by this formula is 98.

For example, an average normalized deviation of .02 produces round(98 − 10.4), or 88. A deviation of .02 represents 2% of landmark-based face width; it does not mean a face is 98% symmetrical. For numerical consistency with the homepage, the implementation sums raw deviations before dividing once by six times face width.

This score measures a narrow geometric relationship. Lens perspective, head angle, expression, and landmark detection can all move the result.

What Your Symmetry Score Means

These descriptive bands explain the formula’s output. They are not population percentiles or attractiveness categories.

90 and above

The measured pairs have relatively small average horizontal differences around the reference line.

75–89

Some horizontal mismatch is present. Check head position and expression before comparing results.

60–74

The pairs have a larger average difference. Review the mirrors and try another straight-on photo.

40–59

The rule detects substantial mismatch. Inspect pose, framing, obstruction, and landmark placement.

Below 40

The measured imbalance is large relative to face width. Check the photo and detection quality before drawing conclusions.

We do not have a representative population dataset, so we cannot say where most people score. Mirrored comparisons deliberately exaggerate differences by repeating one side across both halves. They are visual aids rather than predictions of an ideal face.

Does Symmetry Mean Attractiveness?

A symmetry score cannot determine attractiveness. This test does not measure expression, styling, personality, or how a face appears in movement. It also does not perform a golden ratio face assessment: proportion and symmetry describe different relationships.

A face can have balanced landmarks with many different proportions or face shapes. Use this tool to understand the photograph. Small visible differences do not establish a problem that needs correcting. For a broader geometric breakdown, try the free Face Rater.

Take photos you can compare

Place the camera at eye level and look straight into the lens, keeping both ears similarly visible. Use soft, even light from the front rather than a bright window on one side. Relax your eyebrows and mouth, avoid smiling, and remove filters or portrait effects. Keep the camera distance, framing, and lighting consistent between attempts; avoid close-up wide-angle selfies. If two results differ, compare the original photos before interpreting the numbers. Repeating the test under similar conditions is more useful than selecting whichever image produces the highest score.

Free Score vs. the Full Report

The paid option is the existing comprehensive report, not a five-region symmetry assessment. Percentiles are model heuristics, not measured population rankings.

One-time payment. No subscription or account required. Seven-day refund policy. Start with the Face Rater to purchase.