Jawline Rating: Free Jawline Analysis from a Front Photo

What your result shows, and what it leaves out

Most tools that offer a jawline rating end with a single number. This one does not, and that is deliberate. A front photo can measure how wide the jaw is, how wide the chin is and what angle the jaw outline makes when projected flat onto the image. No published reference exists for what that projected angle “should” be, so any score built on it would be a number we made up. Instead, the card gives you measurements, a plain category and a list of what cannot be known from this photo.

The card also shows a confidence value with one decimal, built from head pose, light and face size, and plain notes when a photo is dark, low in contrast or small. If the head is turned more than 10° or tilted outside the allowed pitch window, the card asks for a retake and withholds the values, because a tilted jaw changes the projected angle more than a real difference between faces would. The annotated copy of your photo marks the jaw corners and chin with red dots; it is drawn on your device and never uploaded.

Why this page gives no number

Scores need a reference, and an honest jawline rating needs one too. The thirds, fifths and symmetry parts of the face rating each compare a measurement with a published or geometric target, and the home page lists every constant. For a front-view jaw there is nothing equivalent. Orthodontic literature measures jaw angles on side-view X-rays (cephalograms), where the gonial angle is a real, visible angle. What a front photo gives is a different quantity: the angle your jaw outline makes when a three-dimensional shape is flattened onto the image. We found no study that defines a norm for that projected angle, and we will not borrow figures for the gonial angle from clinic pages that do not cite a primary source.

What the category thresholds are. Angular is a projected angle below 128°, balanced is 128–136°, and soft is above 136°. They are centred on MediaPipe’s canonical average face model, which measures 132° with the same landmarks. They are labels for where a photo falls relative to that model, not targets, and not a health or beauty standard.

The one peer-reviewed study we cite, Moradinejad et al. (2022), is about how orthodontists, surgeons and laypeople judged female jaw angles on frontal and three-quarter views. It shows that jaw width and angle are studied as matters of perception and preference, which is one more reason not to turn them into a pass mark.

What is measured, and where

Every length is divided by the distance between your pupils (iris centres, landmarks 468 and 473), so the size of the photo does not matter. The eye line is levelled first, and a small head turn is undone using the model’s depth estimate. The jaw measurements are:

The corner points are the model’s estimate of the lowest, widest part of the jaw outline. They are a proxy for the gonion, the bony corner of the jaw, not the bone itself. Where soft tissue, a beard or a shadow hides the outline, the model still places a point, so the outline it traces is the one the photo shows, not necessarily the one under it.

Categories, and what needs a side photo

The category comes from the projected jaw angle alone. The other values are shown as context and are not part of the decision.

The gonial angle is the angle between the lower border of the jaw and the rear edge of the jaw bone, and it lies in the side plane. Chin projection is how far the chin sits forward of the rest of the face, which is a depth, and a photo taken from the front compresses depth to almost nothing. The same limit applies to the nasofrontal and nasolabial angles and to facial convexity, which also need a profile. We say so rather than guess: a value that looks precise but cannot be measured is worse than an empty row.

How to read the numbers

Read the category last when you read a jawline rating from any photo. Start with the confidence value: if it is low, the photo is the likely reason for any surprising number. Then look at the three values together, since they describe different things.

Camera height and head tilt

A camera held below the chin shows the jaw from underneath, which makes the outline look wider and lower and tends to flatten the angle at the corner; a camera held above the face does the opposite. A chin tucked down or lifted up changes the angle in the same way, which is why the tool accepts only a limited pitch range. Small differences in tilt can be enough to move a photo from one category to its neighbour, so a result close to 128° or 136° is best treated as borderline.

Lens distance

A phone held at arm’s length or closer enlarges the parts of the face nearest the lens and shrinks the edges. The jaw outline sits farther back than the cheekbones in a front view, so close-range photos tend to make jaw ÷ face width read smaller. Stepping back and cropping gives a truer width ratio.

Beard, hair and light

The model follows the visible outline. A beard, long hair over the jaw, a high collar or a shadow below the jaw line changes where the contour points land, and so changes the measured width, the angle and the chin values. Hard side light darkens one half of the jaw and can move the points on that side, which also reads as asymmetry on the symmetry analysis.

Width and angle also move together: a jaw that is wide relative to the face usually has a corner that sits lower and further out, so a higher jaw ÷ face width and a lower projected angle often appear together. They are still separate measurements, and the table above shows which landmarks each one uses.

Photo technique, and optional notes

The first four cards change how reliably the jaw is measured. The last two are optional and about preference only.

The measurements above are one view of a three-dimensional shape. For the rest of the picture, the face shape detector uses jaw width and angle with three other features, the face rating scores symmetry, thirds, fifths and golden ratio, and the golden ratio calculator lists each proportion against φ.