Facial thirds, without inventing a hairline

The classical idea of equal thirds is an artistic and clinical reference, not a law of attractive faces. Natural variation is expected.

MediaPipe provides stable landmarks around the brows, nose and chin but not a dependable trichion—the anatomical hairline point. Guessing through hair and presenting the result as measured would add false precision.

• Camera above or below the face
• Chin tipped up or tucked down
• A smile that changes the lower face
• A very close lens
• Cropped chin or brow landmarks

Artists, clinicians and photographers use horizontal guides to organize a complex face. The traditional boundaries—hairline, brow region, nose base and chin—provide a shared vocabulary for discussing vertical proportion. They do not imply that every balanced face has identical segment lengths or that a deviation requires correction.

Clinical uses also depend on purpose and method. A clinician may select anatomical points, imaging and population references appropriate to a functional question. An illustrator may simplify the same face for construction. A browser landmark model has a narrower task and should not borrow clinical authority for an approximate photo ratio.

• Artistic guide: helps construct and compare forms.
• Photographic guide: helps notice pose and crop.
• Clinical analysis: selected for a specific diagnostic purpose.
• Browser estimate: describes visible landmark relationships.
• Beauty claim: requires evidence beyond a traditional diagram.

The trichion is the point where the forehead meets the hairline, but hairlines vary naturally and can be obscured by styling, fringe, coverings, recession or transplantation. A face-mesh model tracks skin and feature landmarks; it does not have a dependable anatomical hairline point. Estimating one from visible hair would confuse styling with anatomy.

Some diagrams substitute the top of the skull or image crop, creating a different measurement while retaining the familiar “thirds” label. Omogle avoids that substitution. It compares the observable middle and lower segments, says exactly which boundaries it uses and leaves the unavailable upper segment unscored.

• Hair can cover the forehead boundary.
• A styled edge may not equal the anatomical hairline.
• The image may crop the top of the head.
• Recession changes the visible boundary over time.
• Inventing a point creates false precision.

The lower segment includes the area from the nose base to the chin and is affected by mouth position, lip posture, dental relationships and chin visibility. Smiling lengthens and reshapes some distances; tucking or lifting the chin changes perspective. A neutral expression and level camera make repeat measurements more interpretable.

Even a stable difference does not establish a universal ideal. Vertical balance interacts with face width, feature size, age, sex-linked averages and individual identity. Use the output to understand the tool’s selected proportions, not to diagnose facial growth or decide that one segment must be changed.

Before comparing diagrams, write down each boundary. Some place the upper edge at the hairline, some at a skull estimate and some quietly use the image crop. Brow landmarks may refer to the eyebrow hair or bony ridge. Chin points may follow skin or imaging. These choices can change the segment lengths while the face stays identical.

Next check pose. Camera height and head pitch alter vertical projection, and smiling changes the lower segment. A side view, front view and radiograph are not interchangeable. Reputable analysis names its construction rather than presenting an unlabeled set of horizontal lines as universal truth.

Omogle’s narrower middle-to-lower balance is intentional. It sacrifices the familiar complete diagram to avoid fabricating an upper boundary. That trade makes the result less grand and more defensible, while preserving the landmarks that a clear, neutral, front-facing photograph can actually support with reasonable consistency.