Make-up semi-automatically!

In fact, you can even create your own makeup look using an interactive evolutionary computation system. The principle is the same as that of men's and women's watches; by combining makeup components such as foundation, blush, and eyeshadow, an interactive evolutionary computation system can generate a variety of unique makeup patterns. Here is an illustration showing the makeup components and examples of makeup looks.


Makeup Component Designs


Makeup Examples *The face images are generated by a GAN and depict fictional characters.

In this system, two virtual makeup patterns are presented on a computer display, applied to a digital version of the user's face. The user then inputs their preference for one of the patterns or indicates a tie. So far, this process is similar to the ladies' watch design. Additionary, in this system, users can specify their preferred design components during the evaluation and experiment with different scenarios. The ability to specify preferred design components (direct manipulation) was introduced based on the concept that "users apply their own makeup," if they only repeat liking or disliking evaluations, the experience of applying makeup feels diminished."


Interface for Direct Manipulation

In addition, makeup trends can vary depending on the occasion, whether for going out with frends or keeping it formal in business scene. Taking this into account, we prepared three scenarios for the experiment: job hunting, going out, and casual makeup, allowing participants to create looks for each situation.

By constructing this makeup system and conducting a series of experiments, we are analyzing its strengths and weaknesses to progress to the next stage.

Related our works
[International Conference] Hiroshi Takenouchi, Shion Isayama, Masataka Tokumaru, "Automatic Virtual Makeup System Using User?Preference Information", HCI International 2023 Posters, Communications in Computer and Information Science, Vol. 1836, Springer, pp. 523-529, 2023-07 (Copenhagen, Denmark).
[Domestic conference] Shion Isayama, Hiroshi Takenouchi, "Measurement of Gaze Information in Selecting Favorite Makeup"" Joint Symposium of the Japan Society of Kansei Engineering, Life Software, Kansei Workshop, and Jiritsu-kai 2022, A2-3, 2022-12 (Hiroshima)(in Japanese). Presentation Award!
*Bold: Students in this laboratory

▲pagetop