Skintone Finder: Personalized Solution to Makeup Color Selection

The solution provides the best matching face makeup products to individual customers and was developed through the digital transformation of expert knowledge, consumer expectation, and chemical engineering toward a personalized premium service provided remotely.

Article | Special Issue

The style business is not an exception in digital transformation. The cosmetic industry struggles to re-create “data” from its long-standing expertise and knowledge while maintaining the personal attachment to individual customers. Often, accuracy-driven technical solutions do not succeed in this context due to the lack of real understanding of users’ desires. Consumers appreciate emotionally touching experience as well as clear answers from an expert.

In this circumstance, the Color Lab of the Department of Industrial Design led by Prof. Hyeon-Jeong Suk has developed an algorithm that distinguishes very tiny areas in color space, which is easily ignored by the conventional approach to color calibration in imaging science. The research team paid particular attention to how consumers perceive human skin color and their anticipation about their ideal selves through the selection of makeup color.

Although the solution is built upon a large-scale dataset, it started with a one-to-one study with individual customers about their makeup behavior. “The procedure was slow and tedious. However, it was essential for us to understand people’s thoughts and desires. The insights became the blueprint of the whole project,” said Yuchun Yan, a Ph.D. candidate who has been carrying out skin color research since her master’s program.

 

The algorithm was finally developed into a solution through iterative trials and errors and was later named “Skintone Finder.” It operates with a flower-shaped apparatus comprised of six color patches deliberately selected through the psychophysical studies with consumers. Based on the distance with the color patches, the algorithm identifies an answer that defines one’s skin undertone and recommends products, as if an expert of the brand were consulting with each customer.

Caption: A promotion video of the service filmed by Amore Pacific, to whom the research team transferred the patented algorithm of the solution.

The intellectual property of the algorithm and “Skintone Finder^TM” solution is registered domestically and internationally, and the technology was recently transferred from the Color Lab at KAIST to the Amore Pacific. Currently, the solution is optimized for mobile devices and has been launched by Amore Pacific’s luxury brands. The algorithm and commercialization are truly valuable outcomes of the research, but the whole story began from the transformation of wisdom and knowledge into digital data and their subsequent restructuring.
“First, we invested time and effort to observe the optical properties until we discovered the underlying logic between physical and psychological quantities. It sounds counterintuitive, but the digital transformation is more meaningful when it is driven by human wisdom,” said Professor Suk.

 

Related achievements:
• Official promotion video at Company’s YouTube Chanel: https://www.youtube.com/watch?v=nEizRvwI5Ow&t=23s
• Journal article: Yan, Y., Lee, J., Hong, J., & Suk, H. J. (2020, published online) Measuring and describing the discoloration of liquid foundation. Color Research & Application, doi.org/10.1002/col.22584
• Domestic patent and PCT registration and applications:
o Suk, H.J. & Yan, Y. (2020), Color palette used for digital portrait to define facial skin color, method, and apparatus using the same [registration no. 10-2170912].
o Suk, H.J., Yan. Y, & J. Lee, “Method for evaluating discoloration of cosmetic composition using opacity charts” [application no. 10-2019-0160004].
• Technology transfer to Amore Pacific. Exclusive license for two years.