A Personalized Visual Aid for Selections of Appearance Building Products with Long-term Effects

Title of the Paper

A Personalized Visual Aid for Selections of Appearance Building Products with Long-term Effects

Paper Information

• Research Domain: Human-Computer Interaction and E-commerce Decision Support
• Keywords: Appearance enhancement products, personalized visual aid, decision support, virtual trial, long-term effects

Research Background and Issues

• What problems or challenges did the authors identify?

  • Selecting appearance enhancement products (e.g., skincare products, weight loss plans) suitable for long-term use is challenging because their effects typically require sustained usage to become evident.
  • Consumers find it difficult to determine whether a product suits their specific needs based solely on product descriptions or other user reviews.
  • Existing online shopping platforms often present excessive information, making it time-consuming and labor-intensive for users to read and understand product details.

• Why is this issue important?

  • The effectiveness of long-term appearance enhancement products varies from person to person, and choosing the wrong product may lead to financial waste and adverse health effects.
  • Personalized decision support can help consumers efficiently select products that meet their individual needs.

• Research Motivation and Related Work

  • Virtual trial technologies are widely used for products with immediate effects (e.g., cosmetics, clothing), but these technologies are difficult to apply to products requiring long-term use to show results.
  • Previous studies have primarily focused on recommendation systems based on user profiles, contextual information, or image analysis, but lack services that visually present long-term effects.
  • The authors aim to develop a visual aid tool capable of predicting long-term skincare effects and intuitively displaying results to enhance user experience.

Solution

• What methods or solutions did the authors propose?

  • Proposed a computational pipeline that utilizes product efficacy labels, customer ratings, and before-and-after images to predict the long-term effects of appearance enhancement products.
  • Developed SkincareMirror as a research prototype, which analyzes user-uploaded facial images to match products suitable for individual skin issues and predict their effects.

• What is innovative about this solution?

  • Provides personalized visual representations of the long-term efficacy of appearance enhancement products.
  • Highlights problem areas on the user's skin with visual markers, enabling quick identification of focus areas and product effects.
  • Integrates product labels, ratings, and image data to reduce the time users spend switching between different platforms.

• Implementation Steps

  • Information Extraction: Collect product efficacy labels, customer ratings, and before-and-after images for classification.
  • Image Processing Parameter Configuration: Use efficacy-based image processing algorithms to generate predicted outcomes on the user's facial images.
  • Module Design and Implementation: Display analysis results as facial image effects to provide intuitive product selection support for users.

Research Outcomes

• What specific outcomes were achieved?

  • Experiments showed that online shopping websites using SkincareMirror significantly outperformed baseline websites in terms of usability, information richness, user satisfaction, and perceived effectiveness.
  • SkincareMirror significantly reduced the time users spent selecting products while enabling them to browse more options.

• What advantages does it have compared to existing solutions?

  • Offers personalized visual predictions, effectively reducing the burden of analyzing traditional textual information.
  • Helps users intuitively differentiate the effects of similar products, particularly for those unfamiliar with the product domain.
  • Displays personalized effects based on user-uploaded photos, significantly enhancing trust in visual outcomes.

• What were the experimental or evaluation results?

  • Participants rated the website's usability significantly higher than the baseline group, and considered the visual effects more trustworthy than consumer-provided images.
  • Female users were more inclined to reference additional product information, while male users relied more on SkincareMirror's predictions.
  • Users unfamiliar with skincare products showed a stronger reliance on the visual aid tool for product selection, with a notable increase in confidence.

• Limitations and Future Directions

  • For products lacking before-and-after images, the system used average parameter values, which may result in less distinct efficacy differentiation.
  • Long-term user satisfaction across different demographic groups and customization for diverse cultural backgrounds have yet to be explored.
  • Issues such as user privacy (uploading personal facial images) and sociocultural concerns in cosmetics (e.g., whitening standards) require further investigation.
  • Future work could explore offline applications of the tool, expand its scope to other domains (e.g., weight loss plans), and optimize the quality of training data.