Skin type Algorithm
What the Skin type Algorithm determines, the classes it returns, and why a reading can change over the course of a day.
About Algorithm
Disclaimer about natural skin behaviour that might affect results Stability and Reproducibility:
Skin Type is a skin health parameter, that normally is not subject to vary on a scale of months. However, natural skin’s sebum production and moisture levels may fluctuate during the day, and it may result in a change of detected Skin Type.
For more precise measurement, we recommend prompting your users to take several pictures within one day and aggregate results to exclude such variations.
This Algorithm estimates the skin type class of the facial skin. Skin Type of skin is a complex state of the skin that is determined by the functional properties of the skin. Skin type is crucial for the recommendations of skin care products that would be beneficial for precise skin characteristic.
This Algorithm estimates indicated skin health parameters on: the T-zone of the face (forehead with the upper part of the nose) and the U-zone of the face (right cheek, left cheek, and chin areas combined). The nose area is not used for evaluation.
The Main Metric for this algorithm is the Skin Type class:
Skin Type class has 4 possible values: "oily", "normal", "dry", and "combination".
These values are based on the market-based classification of skin types. This classification is subjectively functional, based on the client's sensations about the skin's function (amount of sebum secretion, dryness, tendency to certain types of diseases, etc.).
"combination" value is returned for face skin when U-zone and T-zone submetrics are different.
There are two submetrics for this algorithm:
U-zone: one of 3 classes - oily, normal, and dry;
T-zones: one of 3 classes - oily, normal, and dry.
Algorithm API
API access is available only for Professional plan clients
Not a Haut.AI client yet?
Last updated