“It is hard to remove from my eye”: Design Makeup Residue Visualization System for Chinese Traditional Opera (Xiqu) Performers

Title of the Paper

“It Is Hard to Remove from My Eye”: Design Makeup Residue Visualization System for Chinese Traditional Opera (Xiqu) Performers

Paper Information

• Subject Area: Human-Computer Interaction (HCI) and Computer-Aided Skincare Technology
• Keywords: Chinese Traditional Opera, Eye Makeup Residue Visualization, Computer Vision, Mobile Computing, Cultural Heritage, Interactive Design, Skincare System

Research Background and Issues

• Problems and Challenges:
  1. Prolonged use of heavy metal-based greasepaint by opera performers leads to severe skin issues.
  2. Incomplete makeup removal, especially of eye makeup, is a primary cause.
  3. The heavy use of oil-based stage makeup, uneven skin texture, and inadequate cleaning exacerbate the difficulty of removal.
• Significance: Chinese traditional opera (Xiqu), recognized as an important intangible cultural heritage by UNESCO, relies heavily on performers for its preservation. However, skin diseases caused by makeup removal issues affect performers' health and careers, potentially hindering the transmission of this cultural heritage.
• Research Motivation and Related Work:
  • While research in chemistry and medicine has focused on reducing cosmetic toxicity, there is a lack of human-centered skincare assistance technologies.
  • Current HCI studies primarily focus on diagnosing skin diseases rather than preventing skin damage based on performers' needs.
  • Existing technologies, such as wearable devices and augmented vision methods, fulfill certain functions but fail to meet the specific needs of opera performers, such as observing makeup residue or tracking makeup duration.

Solution

• Proposed Method:
  • Designed the EyeVis system, comprising hardware components (fill light, camera lens magnifier, eye mask) and software components (a mobile app-based makeup residue visualization tool).
  • Key functionalities of EyeVis:
    1. Visualizing makeup residue, focusing on the eye makeup area.
    2. Tracking and recording makeup duration.
    3. Enhancing residue detection through algorithms, such as HSV color transformation and binary thresholding.
• Innovations:
  • The first tool dedicated to visualizing eye makeup residue, enabling performers to remove makeup more effectively.
  • Utilizes image processing techniques (non-deep learning), making it compatible with regular mobile devices.
  • Portable, low-cost, and user-friendly design, highly aligned with the daily routines of opera performers.
• Implementation Steps and Techniques:
  1. Hardware Design:
     • Use of fill light to ensure consistent lighting.
     • Lens magnifier to improve image clarity.
     • Eye mask to block ambient light and maintain a fixed distance between the device and the face.
  2. Software and Algorithms:
     • Developed a mobile-based image capture and processing workflow, including basic localization algorithms (e.g., Google MediaPipe) and residue visualization (HSV filtering and binary thresholding).
     • The system records makeup and removal times, generating trend graphs.
     • Proposed a dual-path visualization scheme to highlight residue in different colors.

Research Outcomes

• Experiments and Evaluation:
  • A 7-day deployment study with 12 opera performers validated the usability and effectiveness of EyeVis.
  • Data showed that EyeVis provided stable image quality under various lighting conditions, and its algorithms accurately identified residue areas (average algorithm overlap rate >80%).
  • User feedback indicated that the system not only improved the thoroughness of makeup removal but also raised skincare awareness among users.
• Comparison with Existing Solutions:
  • More focused on solving a single issue (eye makeup residue).
  • The system is more portable and cost-effective, suitable for the highly mobile opera stage environment.
• Limitations and Future Directions:
  1. Data scarcity: Expanding the dataset (including more opera roles and different makeup colors) could enhance algorithm robustness.
  2. Personalized design: Adapting to different users' facial features.
  3. Application Expansion:
     • Extending residue detection from the eye area to the entire face.
     • Broadening the target audience to include general cosmetic users.
     • Serving as a skin data collection tool to support broader skin health research and clinical diagnosis.

This paper integrates HCI with traditional culture, offering a unique solution that addresses practical issues while contributing a technological perspective to cultural heritage preservation.