Creating Aromatherapy Insight Cards using Generative AI: Focused on Midjourney Image to Image
This study analyzed the visual expression characteristics of aroma essential oil images using generative AI, and compared the differences between AI-generated aroma cards and existing aromatherapy insight card images to derive important implications for the development of visual content. The study utilized the generative AI tool Midjourney and the communication platform Discord, and the researcher reconstructed and used prompts based on the subject, medium, and environment when creating the image. The results of the study showed that generative AI effectively visualizes the delicate and luxurious characteristics of aroma essential oils by utilizing harmonious colors and stable compositions centered on female figures. Even for images with the same aroma plant as the theme, AI and existing methods conveyed different moods and meanings through different visual expressions, and the AI-generated aroma card images received high marks for their realistic and creative texture and composition. Expert evaluations clearly showed that the AI-generated aroma card images were significantly superior to existing aromatherapy insight card images in all categories, including creativity, expression of colors and composition, visual appeal, suitability, and value. This study suggests that generative AI is a very effective tool for creating visual content related to aromatherapy, and suggests that it is necessary to seek more practical and specific ways to utilize it in the future through localization that reflects cultural backgrounds and user response surveys.
Authors
- Yun-Kyoung Oh (ORCID: https://orcid.org/0000-0003-3316-9462)
- Jin‐Young Choi (ORCID: https://orcid.org/0000-0003-2676-3842)
- Kyoung-Eun Lee
Publication Details
- Journal
- Journal of the Korean Society of Cosmetology
- Published
- 2025-08-28
- DOI
- https://doi.org/10.52660/jksc.2025.31.4.833
- Primary Topic
- Diverse Approaches in Healthcare and Education Studies
- Type
- article
- Field-Weighted Citation Impact
- 0.00