Uniformness Algorithm (experimental)
What the experimental Uniformness Algorithm estimates, the metrics it returns, and why foundation and other cosmetics affect it.
About Algorithm
Disclaimer
The Uniformness Algorithm is experimental. Results are approximate, may vary under different conditions, and are not intended for medical diagnosis or treatment.
Skin uniformness is not a scientifically defined parameter, as many factors and pigments can influence the skin’s visual appearance. Visual uniformness may differ significantly from natural skin uniformness when cosmetics such as foundation, concealer, BB/CC creams, color correctors, or setting powders are applied before taking a photo.
The algorithm is sensitive to image quality and lighting. Insufficient spatial resolution (below 2500 pixels), uneven lighting, shadows, or reflections can affect detection and cause fluctuations in the uniformness score.
This Algorithm estimates the visible skin texture non-uniformness, such as eruptions, age spots, freckles, and blood vessels close to the surface, along with texture-associated skin features.
This Algorithm estimates indicated skin health parameters on: forehead, right cheek, left cheek areas.
The Main Metric for this Algorithm is the Uniformness Score, with a range from 0 to 100.
Values from 80 to 100 are considered a relatively good skin condition with relatively uniform texture skin.
Values from 50 to 70 are considered as skin with visible skin texture non-uniformness.
Values below 50 are considered as an alerting skin condition, that might require specialist attention.
Uniformness Score is a non-linear transformation of a spatial information about non-uniform skin texture clusters.
There are no Submetrics for this Algorithm.
Algorithm API
API access is available only for Professional plan clients
See Uniformness Result Scheme.
Not a Haut.AI client yet?
Last updated