Matching Virtual Makeup Try-Ons to Different Skin Tones

When your potential customer is shopping online for beauty products, how do you know which foundation suits their face?

Matching Virtual Makeup Try-Ons to Different Skin Tones

When your potential customer is shopping online for beauty products, how do you know which foundation suits their face? Skin tone-inclusive makeup is one of the biggest challenges in the cosmetics industry nowadays.

When the shopper is dissatisfied with a product, most brands have return policies. The returned products are usually tossed away for sanitary reasons since the article has already been opened and used. On the customer’s end, the experience leaves them frustrated, swearing to never buy another single product online. Which is a lose-lose for both parties. A total waste!

So how do you make sure your clients are getting the most accurate pigment for their skin, and consequently maximize your sales? Designhubz has all the answers.

How do we classify skin tone?

There are 2 main references for skin tone. It started with Von Luschan’s chromatic scale that was later replaced by the Fitzpatrick scale. Von Luschan’s schema was used to classify races in populations worldwide. While Fitzpatrick’s scale was to study tanning habits. As Designhubz, we use Fitzpatrick’s as the basis for skin tone calibration. It has less bias, while being a solid foundation to accommodate as much of the spectrum as possible. The interesting part is that we can and are planning to accommodate any and all shades (and everything in-between) using deep learning as we gather more sample data from our users.

How does deep learning work in this case?

Deep learning analyzes a user’s skin - not just their tone – and combines it with all the data collected from them, and similar skinned users’ behavior, in order to create more nuanced scales in between the existing ones. The result will create a more personalized recommendation rather than lumping people with comparable skin color.

What is it used for?

1. Hyper personalized recommendations: we factor in each user’s skin tone to find the best complementing or contrasting makeup color. This feature can also be used for other items such as eyewear.

2.Accuracy in product depiction: our software blends the skin shade with the makeup to display a representation as accurate as currently possible of what the applied product will look like, while factoring in environmental elements such as lighting. It’s crucial to mention that its capabilities are always improving!

What kind of lighting is optimal when shopping online for makeup?

Indirect or natural lighting are best to capture your user’s face. For example if they were facing the sun, the skin would look lighter/paler, whereas skin looks darker in a dimly lit room. Another important aspect is for the shopper to show their actual appearance on camera. And that means no filters!

What does our future hold?

Future features could include pore detection, reflectivity index, skin type (oily, combination, dry etc.) and fine line presence to recommend suitable products. Our goal is to depict the applied product even more realistically on human skin. We aim to create the most realistic online experience in the industry of online makeup shopping.

If you want to find out more about our product or book a demo, check our Makeup Try-on solution and give it a try!

Chady Karlitch

Chady Karlitch

Designhubz Co-Founder & CTO