Face length to cheekbone width
We measure a straight line from the upper facial contour to the chin, then divide it by the distance between the outer cheek landmarks. A ratio of 1.4 means the visible length is about 1.4 times the width. It does not mean the face has an ideal proportion. The upper contour is a model estimate, not a verified hairline.
Cheekbone width and jaw width
The cheekbone-to-jaw ratio describes how much the lower face narrows. Its reciprocal, jaw-to-cheekbone width, expresses the same relationship in the other direction. A jaw-to-cheek ratio of 0.8 means the selected jaw landmarks are about 80% as far apart as the cheek landmarks.
Estimated upper-face width
We compare the temple-region landmarks to cheekbone width. This is labeled “estimated upper face” throughout the report. It does not measure the hairline, forehead height or the full bony forehead. Hair and facial contours may make this estimate less reliable.
Lower-face contour
We calculate a two-dimensional angle at each selected jaw point using the neighboring outline landmarks, then average the two sides. This contributes a small amount to the shape score. It is a contour angle, not a clinical jaw or gonial angle.
A limited symmetry indicator
When head yaw and pitch are within five degrees, we reflect paired landmarks across the upper-face-to-chin line and measure their average mismatch as a fraction of cheekbone width. This describes the photograph, including perspective and model error. It is unavailable for less frontal photos. Natural asymmetry is ordinary.
How to read the charts
The bars show the relative magnitude of each measured ratio on a labeled display scale. Their endpoints are chart bounds, not desirable limits. Match-score bars show agreement with our heuristic shape prototypes, not probabilities or confidence intervals.
Where to go next
Read the face shape guide for how these features combine, or learn about the analysis method and its limitations.
Explore my proportions ↗