FWHR: How to Measure Facial Width-to-Height Ratio

FWHR means facial width-to-height ratio. You calculate it by dividing the face's bizygomatic width, from one cheekbone landmark to the other, by a defined upper-face height. The catch is “defined.” Studies use different top landmarks, so two careful measurements can yield different numbers. Pick one published method, record its landmarks, and compare only results made with that method.

What FWHR actually measures

FWHR is a unitless proportion, not a face size. If width is 420 pixels and height is 250 pixels, the ratio is 1.68. Millimeters would give the same ratio if both distances came from the same flat plane and scale.

The width runs between the left and right zygion points, near the widest visible cheekbone area. The height covers part of the upper face. It does not include the jaw or chin.

That makes FWHR different from a face-shape label. Round, oval, square, and oblong describe the broader outline. Our guide to face shape versus facial proportions explains which question each system answers. If your goal is an outline estimate for hairstyle planning, use the on-device face-shape finder instead.

Three FWHR methods use different height lines

There isn't one universal photo protocol. Carré and McCormick's 2008 study used the distance from the upper lip to the brow. Kramer, Jones, and Ward used the upper eyelids in their 2012 photo method. Their direct facial measurements used the nasion, the depression between the eyes, as the top point.

After leveling the eyes, width is the horizontal difference between the two zygion points. It isn't a diagonal line between uneven points. The denominator is the chosen vertical height, not the full height of the face.

Common facial width-to-height ratio definitions

Method	Width landmarks	Height denominator	Useful for	Main limit
Brow photo method	Left zygion to right zygion	Upper lip to brow	Repeating that exact photo protocol	Brow position changes with expression
Eyelid photo method	Left zygion to right zygion	Upper lip to highest upper-eyelid points	Repeating the Kramer photo protocol	Eyelid and lip points still depend on expression
Direct anthropometry example	Left zygion to right zygion	Upper lip to nasion	Controlled direct facial measurement	The nasion is hard to locate on a casual 2D photo

Kramer's photo protocol averaged the vertical positions of the two highest eyelid points. The height then ran from that averaged line to the highest point of the upper lip. A brow-to-lip height uses a different top line, even when the width is identical.

These methods can correlate without producing matching values. In Kramer and colleagues' controlled sample of 155 White British students, 2D and 3D ratios correlated at r = 0.81. That agreement does not make their landmark definitions interchangeable.

How to measure FWHR from a photo

This is an editorial home method, not a validated anatomical exam. It can produce a repeatable photo estimate if you control the inputs.

1. Choose the brow or eyelid protocol before placing any landmarks. Write the choice beside the result.
2. Use a straight-on photo with a neutral, closed-mouth expression. Keep the camera, lens setting, distance, and lighting consistent.
3. Rotate the whole image until the pupils sit on one horizontal line. Do not stretch or warp the face.
4. Mark both visible zygion points, the upper-lip point, and the selected brow or eyelid line.
5. Measure both distances in pixels, then divide width by height. Save the uncropped photo and landmark definition.

The hardest step is usually finding the zygions. Hair, cheek fullness, lighting, and a small head turn can hide them. If either side is unclear, report the result as uncertain rather than moving the point until the number looks right.

Illustrative arithmetic: one width, two answers

Illustrative example. The coordinates below are arbitrary pixels on a synthetic grid. They are not measurements of a person, a product result, or a population norm.

How the chosen height denominator changes FWHR

Measurement	First landmark	Second landmark	Distance	Ratio calculation	Result
Shared width	Left zygion x = 200 px	Right zygion x = 620 px	420 px	Not applicable	Not applicable
Method A height	Upper eyelid y = 310 px	Upper lip y = 560 px	250 px	420 px ÷ 250 px	1.68 unitless
Method B height	Brow y = 270 px	Upper lip y = 560 px	290 px	420 px ÷ 290 px	1.45 unitless

The face width never changed. Moving only the top height landmark changed the ratio by 0.23. Treating 1.68 as a meaningfully higher FWHR would be invalid because the two values answer different geometric questions.

Why the same face can give a different FWHR

A photo records projection, pose, and expression. It does not expose bone dimensions directly. Robin Kramer tested 21 White adult men at seven camera distance and focal-length combinations in 2016. Distance shifted FWHR, although differences between faces explained more variation.

Expression mattered more in another part of that paper. Kramer compared eight expressions in 20 White adult men. Expression explained more variation than identity, while happy and disgusted poses increased the measured ratio. Fearful and surprised poses lowered it.

Those samples were small and included only White men, so they don't define universal correction factors. They do show why a smile measurement should not be compared with a neutral portrait. Head tilt, left-right rotation, and hidden cheek landmarks add more uncertainty.

If you want to compare two photos, match the capture conditions and use the same landmark protocol. Our guide to measuring facial proportions from a photo covers pose, lens, and landmark limits in more detail.

What an FWHR number cannot tell you

FWHR does not reveal personality, aggression, dominance, ethnicity, health, or personal worth. A single ratio also cannot establish attractiveness. Those claims turn a noisy group-level research measure into an unsupported judgment about one person.

It isn't a medical measurement or a diagnosis. Soft tissue, expression, body composition, and camera setup affect the visible result. A photo estimate also can't separate cheekbone width from the tissue over it.

There is no responsible “ideal” number to chase across unrelated calculators. First ask which landmarks a result used. If the method is missing, don't compare it with your measurement. The useful next step is simple: repeat one disclosed protocol on one controlled, neutral photo and label the result with that protocol.