Amorepacific Unveils AI Virtual Makeup, Skin Prediction, and Aging Microbiome Findings at IFSCC Congress 2025

Amorepacific presents AI-based skin science innovations at IFSCC Congress 2025

At the 35th International Federation of Societies of Cosmetic Chemists (IFSCC) Congress in Cannes (Sept. 15-18), Amorepacific shared AI-driven virtual makeup systems, a skin prediction pipeline with clinical outcomes, and an integrated microbiome-metabolite analysis for aging research. For R&D teams, the throughline is clear: AI methodology is moving into validated practice, while multi-omics is opening new segmentation paths in skin aging.

Generative AI architecture for virtual makeup

Myeongjin Goh (Amorepacific R&I Center) presented "Development of the Generative AI-based Pipeline Architecture for Advanced Virtual Makeup," co-developed with Korea Advanced Institute of Science & Technology. The system applies generative models and image synthesis methods specific to makeup, analyzes facial shapes and features, and uses extensive makeup trend data to recommend personalized styles. Users can preview realistic, AI-recommended colors and looks before any physical application.

• R&D considerations: ensure accurate face geometry analysis, stable color rendering under varied lighting, and evaluation against perceptual realism metrics.
• Data focus: coverage of skin tones, textures, and style preferences to minimize bias and improve recommendation quality.

Skin prediction algorithm with clinical outcomes

Eunbi Ko presented "Application of a Skin Prediction Algorithm for Personalized Skincare Services with Clinical Outcomes," validated at Amorepacific's in-house City Lab. Building on the "Future Skin Prediction Algorithm" and "Customized Skincare Solution" introduced at the 2022 IFSCC Congress in London, the work showed measurable improvements via AI-based prediction linked to personalized programs. Results drew on real customer reviews and controlled clinical trials.

• Implementation notes: couple prediction with intervention, track real-world feedback, and align clinical endpoints with model targets.
• Operational loop: periodic model updates using longitudinal outcomes to improve individualization over time.

Integrated microbiome and metabolite analysis for aging

In the poster "Integrated Analysis of the Age-related Microbiome and Metabolites in Korean Women's Skin," Kil Sun Myoung reported an integrated view across microbiomes, skin physiology, and metabolites by age group. The team identified age-specific traits and, notably, found that microbiome profiles among women in their 20s and 60s split into three distinct clusters, not two. This points to heterogeneous aging pathways that simple age brackets may miss.

• Implications: consider stratified study designs that reflect multi-cluster aging patterns rather than broad age bins.
• Next step: validate clusters longitudinally and link them to functional biomarkers and intervention response.

Participation at the Congress

Established in 1959, the IFSCC Congress is the largest conference dedicated to cosmetic science research. The 2025 theme, "The Future is Science," framed Amorepacific's two oral and five poster presentations advancing AI integration and biological aging research. See event details via the IFSCC.

Strategy: AI First and Ageless

"Amorepacific is advancing toward its 2035 vision, centered around the company's 'Ageless' and 'AI First' strategies. The Amorepacific R&I Center will deeply integrate AI technology across all areas of research and development to expand the boundaries of skin science and set new standards for future innovative beauty," said Byung-Fhy Brian Suh, CTO and Head of Amorepacific R&I Center.

What R&D teams should watch

• Standardized evaluation for makeup simulation (color accuracy, geometry fidelity, user acceptability).
• Bias checks across skin tones, ages, and lighting conditions; clear reporting on dataset composition.
• Privacy-by-design in imaging pipelines and clinical workflows.
• Interpretable models for skin prediction, connecting features to biological mechanisms.
• Multi-omics frameworks that tie microbiome and metabolite shifts to phenotype and treatment response.
• Clinical study designs that close the loop from prediction to intervention to outcome.

Further reading

Explore the IFSCC community and resources at the official site: ifscc.org. For teams building similar AI workflows, you can review curated training on applied generative modeling and evaluation here: Complete AI Training - latest AI courses.