∆/VƏR∆GƏ
The Metron of Deviance
if the face is a canon, the deviation is the self
The average face is considered beautiful. In 1990, Judith Langlois and Lori Roggman averaged photographs of faces by computer and found the composites judged more attractive than most of the faces they were made from. A statistical observation became an aesthetic law.
∆/VƏR∆GƏ measures your face against the canon of facial proportion — Vitruvius's thirds, the artists' rule of fifths, the neoclassical canons, and the golden ratio that modern beauty culture added to them. It returns a deviation index. Then it inverts the verdict: the closer you are to the canon, the more average you are.
The system is not measuring beauty. It is measuring conformity to a mathematical abstraction and calling that conformity beautiful. The canon is not a person. When Farkas and colleagues measured real faces against it in 1985, the canon often failed to fit them. You are being judged against something that does not exist.
Submit your face. Receive your deviation score. The metron does not negotiate.
Ok pretty face. Your image never leaves your device: face detection runs locally in your browser and nothing is stored. If the AI verdict is active, only the measurements — never the image — are sent to Anthropic's Claude API to write the text. Here visits are counted, not people: GoatCounter, open source, no cookies. It sees the page, where you came from and your browser, keeps only the totals, and forgets you.
L∆UNCH ∆PP →