Newly Developed Facial Impression Evaluation Method Utilizing AI and AR Technologies
- Visualizing the ideal facial impression to achieve beauty unique to each individual -

　FineToday Co., Ltd. (Head Office: Minato-ku, Tokyo; Representative Director, President and CEO: Tetsuo Komori) has newly developed a facial impression evaluation methodology utilizing a simulation tool that leverages artificial intelligence (AI) and augmented reality (AR) technologies. Through this initiative, new insights have been gained to help men express their ideal facial impression.
　The findings of this study were presented at the 27th Annual Meeting of Japan Society of Kansei Engineering (September 17 to 19, 2025, Tokyo, Japan) and the 30th JFACE Annual Conference (November 1 to 2, 2025, Tokyo, Japan).

■Background
　Due to the prevalence of social media use and online meetings, it has become increasingly common to see one's own face on screen. Because of this, men have come to show more interest in beauty, leading to the growth of the male cosmetics industry in recent years. There is now a trend especially among younger generations to value their individuality or uniqueness, and thus the way men express themselves has diversified. However, given that men in general have less beauty-related experiences, like putting on makeup, compared to women, the challenge of choosing beauty products that suit them out of a multitude of information and numerous products, as well as the averseness to "overdoing" their looks, is discouraging many men from starting new beauty routines. In addition, in the field of impression research, although attempts have been made to understand consumers' ideals through psychological assessments and other means, it has been difficult to clarify an image that consumers themselves cannot articulate.
　To address these issues, we have developed a new method using a simulation tool independently created with the AI and AR technologies of Perfect Corp. (Minato-ku, Tokyo). This tool can instantly and freely alter a person's facial features to evaluate the impression a viewer gets, and we used it to research the facial impressions men wish to have and the approaches to achieve them via facial transformation.

■Research details
　We conducted the following experiments using the simulation tool with common Japanese men.

[Experiment 1] Investigation into one's own "ideal" and "handsome" faces
Subjects:
63 Japanese men, aged 20 to 49

Method:
The subjects took a photo of their own faces and created the "ideal" and "handsome" versions of them by altering the eyes, pupils, nose, mouth, and eyebrows.

Results:
There was a significant difference in the eye angle, with "handsome" faces tending to have the outer corners of the eyes raised compared to "ideal" faces (Table 1). Furthermore, while there was only a single peak in the distribution of eye angles for "ideal" faces, the distribution was widely spread out for "handsome" faces (Figure 1). A concentration in the distribution suggests the existence of a shared ideal among subjects with respect to their own visage, while a spread indicates that attractiveness has become diversified in society.

Table 1: Comparison of "ideal" and "handsome" faces

Figure 1: Distribution of eye angles

[Experiment 2] Investigation of transformation approaches to create a "handsome" face when editing one's own face and another person's
Subjects:
Self-edit group: 63 Japanese men, aged 20 to 49
Other-edit group: 31 Japanese men, aged 22 to 49

Method:
Subjects were divided into a "self-edit group" that altered their own faces and an "other-edit group" that altered other people's faces. The self-edit group used the simulation tool to freely adjust the eyes, pupils, nose, mouth, and eyebrows of their face photos to create a "handsome" version of their own face. The other-edit group made the same alterations to original face photos of five randomly chosen subjects from the self-edit group, creating a "handsome" version of other people's faces.

Results:
A significant difference was observed between the self-edit group and other-edit group with respect to eye size, distance between the eyes, and width of the bridge of the nose. The self-edit group enlarged their own eyes and narrowed the bridge of their nose to emphasize a pleasant and friendly impression, whereas the other-edit group brought the eyes significantly closer together to give deep-set eyes and a dignified feature, giving prominence to vigor and manliness.

Table 2: Comparison of transformation approach between self-edit group and other-edit group

Figure 2: AI-generated image of a man's face,

altered using the simulation tool by reflecting the transformation characteristics of self-edited and other-edited faces

　This research revealed that there is a difference between the "ideal" and "handsome" faces envisioned by men. Moreover, it became clear that the approach to creating a "handsome" face differs depending on whether a man edits his own face or another person's. These findings indicate that research incorporating the perspectives of both self and other can be valuable for developing solutions that help achieve ideal facial impressions.

■Future prospects
　Based on our Purpose (our reason for existence) to "enrich the lives of everyone today and for generations to come, one fine day at a time," we have developed products with aesthetics as their foundation. Particularly for men, we have had groundbreaking offerings, including "uno," "Ag DEO 24," and others that bring out the unique charm of each man.
To meet the demands of diversifying beauty standards, we will not limit ourselves to a single definition of beauty. And by leveraging the new evaluation method for facial impressions, we will continue our impression research embodying various perspectives, including gender, age, and nationality. By clarifying the ideal impression each consumer holds and the approaches to realizing them, we will strive to develop solutions that support everyone to achieve their unique beauty.

＜Links＞
Official website：https://www.finetoday.com/en/

Show Full PDF (PDF：506.2 KB)