What the PSL score actually measures

Nothing in the scan asks anyone what they think of your face. Every input is a distance, an angle or a ratio taken from your own landmark mesh and normalised to your own face size, so the reading does not change when you move the camera closer.

• Eye region: canthal tilt, spacing, openness, brow distance
• Midface: facial thirds, facial fifths, midface ratio, nasal index
• Lower face: gonial angle, jawline definition, chin to philtrum ratio
• Mouth: lip ratio, lip fullness, mouth to nose width
• Whole face: symmetry, facial width to height ratio, face shape

How the score is built from the readings

Each reading is compared with a geometric reference target and a tolerance band, producing a 0 to 100 sub score. The headline number leans on the readings that stay stable between captures, because a measurement that swings between photos cannot carry a score. The full formula, every target and every tolerance is published on the methodology page.

Why tier labels are not used

Tier language turns a measurement into an identity, and the tiers themselves are not defined by anything measurable. Facemap reports a number, a percentile and the specific readings that drove it, so the result points at geometry you can actually look at rather than a label you cannot escape.

Getting a reading you can trust

Three photographs are taken at different distances and the best quality capture supplies the headline score. Head rotation is the single biggest source of error, so the capture step guides your angle and rejects frames it cannot read properly.

• Face the camera straight on with a level chin
• Even, front facing light, no strong side shadow
• No hat, no glasses, hair clear of the brows
• Relaxed mouth, no wide smile
• No filters or beauty modes