Technology

The science behind Dermaself AI

Dermaself reads visible skin characteristics from a selfie and analyzes the ingredient list (INCI) of every product, so each recommendation matches what the skin needs with what the formula contains.

Contact us
Woman scanning her face with the Dermaself app on a smartphone

How AI Skin Analysis works

From a single selfie to a detailed reading of the skin, in three steps.

Woman taking a selfie for an AI face scan

Take a guided selfie

On-screen guidance helps the user frame the face correctly and take a photo the AI can read.

Illustration of the AI skin analysis process on a selfie

The AI reads the image

Computer vision analyzes the selfie with criteria developed by Derming over 30 years of dermatological research.

Dermaself app screen with the skin analysis chart

The skin profile is ready

The result is a profile of visible skin characteristics, such as blemishes, spots, wrinkles, redness and pores.

What the analysis detects

The analysis measures more than 60 skin parameters. These are the main areas it covers.

8 types of blemishes

Spots

Wrinkles and fine lines

Redness

Pores

Dryness

Under-eye bags

Skin texture

And more

PRODUCT RECOMMENDATIONS

From skin analysis to product recommendations

Skin results are matched with the products in the brand's catalog and with their ingredients, both on e-commerce and in physical stores. The outcome is a personalized routine, not a list of bestsellers.

Complete routines

Products from the catalog are combined into one complete routine for each skin.

Complementary products

Products that work well together appear side by side, opening cross-selling opportunities.

A clear reason for every product

Each choice is based on skin needs and ingredients, so it can always be explained.

Easier purchase decisions

Personalized advice helps shoppers choose the right products with more confidence.

INGREDIENT INTELLIGENCE

How the ingredient list (INCI) is analyzed

Dermaself analyzes the ingredient list (INCI) and the composition of every product in the catalog. Recommendations are based on what a formula contains, not only on marketing claims.

Illustration of the European Commission CosIng cosmetic ingredients database

European Commission CosIng database

Ingredient data comes from CosIng, the European Commission database of cosmetic ingredients.

Source: European Commission – CosIng

Illustration of cosmetic ingredient compatibility

Ingredient compatibility

Compatibility rules check that the products chosen for the same routine have ingredients that work well together.

Illustration of matching skin characteristics with ingredients

Skin and ingredient matching

Each skin characteristic is linked to the ingredients that address it, following criteria from dermatological research.

RESEARCH AND INNOVATION

Developed with Politecnico di Milano and Politecnico di Torino

Two research projects make selfie analysis more reliable and protect user data.

Image quality

Selfie color normalization

Illustration of selfie color normalization

Developed with Politecnico di Milano, color normalization reduces the differences caused by lighting, so selfies taken in different places and conditions can be compared with each other.

Politecnico di Milano

Privacy

Selfie anonymization

Illustration of selfie anonymization

Developed with Politecnico di Torino, this system turns images into irreversible anonymous data that keep their analytical value without being linked to the user. It is designed to comply with the GDPR.

Politecnico di Torino

30 years of dermatological research

Dermaself's AI is built on the work of Derming, a clinical research institute with 30 years of dermatological research. The same scientific base guides both the skin analysis and the product recommendations.

See AI Skin Analysis in action

Book a demo to see how a selfie becomes a skin profile and a personalized routine built from your catalog.

Contact us →

Frequently Asked Questions

Quick answers to common questions

How does AI Skin Analysis work from a selfie?

Dermaself analyzes a selfie with computer vision, the branch of AI that interprets images, and measures more than 60 visible skin parameters such as blemishes, wrinkles, redness and pores. The results are then matched with a brand's catalog to recommend a personalized skincare routine.

How are ingredients used to personalize skincare recommendations?

The ingredient list (INCI) of each product is compared with the skin characteristics found in the analysis. Ingredient data comes from the European Commission's CosIng database, and compatibility rules check that the products in a routine work well together.

Is AI Skin Analysis a medical diagnosis?

No. The analysis describes visible skin characteristics for cosmetic purposes only and does not diagnose or treat any condition. For medical concerns, users should always consult a dermatologist.

Does lighting affect the results?

Lighting changes how skin looks in a photo, so the AI applies color normalization, developed with Politecnico di Milano, to reduce those differences. Selfies taken in different conditions stay comparable, and guided capture helps users take a usable photo in the first place. Read more on AI skin analysis accuracy across phototypes.

How is user privacy protected?

Selfie anonymization, developed with Politecnico di Torino, turns images into irreversible anonymous data that cannot be linked to the user's identity, while keeping their analytical value. The system is designed to comply with the GDPR. See our guide to selfie privacy and anonymisation.

Contact us

Tell us about your brand or project: our team will get back to you by email.

Contact us

Tell us about your brand or project: our team will get back to you by email.