본문 바로가기
  • Home

Development of a Personalized Makeup Recommendation and AR Simulation System Based on Personal Color Analysis

  • Journal of The Korea Society of Computer and Information
  • Abbr : JKSCI
  • 2026, 31(9), pp.85~96
  • Publisher : The Korean Society Of Computer And Information
  • Research Area : Engineering > Computer Science
  • Received : August 5, 2026
  • Accepted : September 4, 2026
  • Published : September 30, 2026

Mi-Young Song 1,  Yong-Sun Kim 1,  Sangmi Im 2

1수원여자대학교
2한국영상대학교

Accredited

ABSTRACT

This study proposes a personalized makeup system that analyzes users’ personal color types based on skin color data obtained from facial images, recommends suitable color makeup, and enables virtual application in an augmented reality (AR) environment. The proposed system detects facial regions and key facial landmarks from input facial images and selects regions of interest (ROI) for stable skin color analysis. The extracted color data are normalized and transformed into CIE L*a*b* based feature values through a color space conversion process, which are then used to quantitatively analyze skin tone characteristics. The analyzed features are utilized to determine the user's personal color type through personal color classification criteria and classification methods, and suitable color makeup is recommended according to the identified type. The recommended makeup colors are applied to facial images through an AR-based synthesis process, allowing users to preview virtual makeup results. This enables users to compare and evaluate makeup applications suitable for their personal color before actual application. This study complements the existing personal color analysis methods that rely on experts’ experience and subjective judgment by introducing a quantitative analysis approach based on image data. Furthermore, it proposes an integrated digital beauty system that connects personal color diagnosis, makeup recommendation, and AR simulation functions. Future research will focus on improving color correction methods to enable stable skin color analysis under various lighting conditions and shooting environments. In addition, by securing diverse user datasets, the performance of personal color classification and the quality of AR makeup application will be further enhanced.

Journal Copyright Policy

No CCL information provided

Citation status

* References for papers published after 2025 are currently being built.