Areum is an aesthetics research lab studying human beauty through open research, beginning with the face. Its work combines 3D facial measurements with judgments from a global panel to investigate how perceptions of attractiveness vary across populations. Rather than presenting beauty as a single universal standard, Areum reports findings through demographic lenses, including differences across ethnic, age, and gender groups.

The lab’s flagship product, Seonho-1, is a machine-learning model trained on millions of pairwise comparisons of real faces. Users can create an account, answer a short set of questions, and complete a guided face scan on the web. The model uses the scan’s measurements to generate a facial rating, percentile, z-score, and demographic breakdown. Reports also communicate uncertainty, and ratings may shift as new votes refine the reference data.

Areum invites people to participate by voting on face comparisons. These judgments contribute to the research dataset, while findings are presented as population-level statistics rather than individual voter data.