Heart Face Shape
Check it here first. Type the three ratios from your own measurements; this page compares them with the heart profile using the same published targets and tolerances as the detector. If you have no numbers yet, use the photo detector — it reads the four distances off a photo and compares all seven profiles, not only this one.
The placeholder values are the heart targets. A ratio may sit up to 0.30 away on the length ratio and up to 0.12 away on the two width ratios and still count as a match.
The heart profile, as published
| Ratio | Heart target | Allowed distance |
|---|---|---|
| Face length ÷ cheekbone width | 1.42 | 0.30 |
| Forehead width ÷ cheekbone width | 1.00 | 0.12 |
| Jaw width ÷ cheekbone width | 0.70 | 0.12 |
Those five numbers are copied from the site's own reference file, face-shape-ratios.csv, row Heart, published at the data page under CC BY 4.0. They are the profile this site compares against; they are not a standard from anywhere else, and you can download the file and argue with them.
How a heart face shape is put together
Four distances are read off the face: forehead width (temple to temple), cheekbone width (the widest point of the face), jaw width (jaw corner to jaw corner) and face length (top of forehead to bottom of chin). Face length counts as 100 by definition, and the other three are reported as percentages of it — so these are proportions inside one photo, not millimetres or inches. The landmark indices each distance is taken between are in the same published set, face-shape-measurements.csv.
Two of the three heart numbers are what make it recognisable: the forehead is as wide as the cheekbones (1.00), and the jaw is clearly narrower than both (0.70).
Measuring a heart face shape by hand
Soft tape, flat against the skin, at arm's length from the mirror. Write numbers in the sheet, not adjectives. The printable tape-measure worksheet (PDF) has a row for each of the four distances, and the step-by-step method is on how to know what face shape you have. Everything in that worksheet and in this page's tables comes from the same published data.
What using this page costs
Nothing, and there is no catch hiding behind the word: there is no charge, no account and no upgrade tier on this site, and there is no daily limit or cap on how many photos or measurements you use — nothing here counts your uses. You do not hand over an email address to read these numbers, and there is no paid version of this page.
Where your photo and your numbers stay
The detector is one page of JavaScript plus a model file. The model downloads once with the page and then the measuring happens inside your browser tab — your photo is never uploaded anywhere, and no server receives it. The comparison box above works the same way: the three numbers you type are compared inside the page and are not sent anywhere.
Heart against the profiles it is confused with
| Profile | Length ÷ cheekbone | Forehead ÷ cheekbone | Jaw ÷ cheekbone | What separates it from heart |
|---|---|---|---|---|
| Oval | 1.45 | 0.86 | 0.76 | Narrower forehead |
| Diamond | 1.50 | 0.74 | 0.72 | Much narrower forehead |
| Square | 1.22 | 0.97 | 0.95 | Jaw nearly as wide as the cheekbones |
| Triangle | 1.32 | 0.80 | 1.02 | Jaw wider than the cheekbones |
All four rows come from the same published file as the heart row. A face can sit close to two of these rows at once — which is expected, because the tolerance on the length ratio is wide.
Questions about the heart face shape
Is a heart face shape the same as an inverted triangle?
This site keeps one row for it: widest at the cheekbones, forehead as wide as the cheekbones, jaw clearly narrower. There is no separate inverted-triangle row, so a face other charts would call an inverted triangle lands in the heart row here.
What if my numbers sit between heart and diamond?
Then both profiles are close on your face. The detector picks the closest of the seven; the box on this page only tells you whether heart fits at all.
Does this page measure my photo?
No. It compares numbers you type in. To measure a photo, use the face shape detector.