University & Competitions

Missing Person Alert Improvement Study

A capstone study that tested wording, layout and content of missing-person alerts through surveys

Outstanding Presentation Paper, SCI2026

Period | 2025.03 ~ 2025.06

  • #Python
  • #Survey Analysis
  • #Kakao Map API
  • #Stable Diffusion
  • #IP-Adapter
  • #InsightFace

Why do people ignore missing-person alerts?

Problem

  • Missing-person text alerts launched nationwide in 2021, but they convey key details poorly and feel distant
  • In the first survey, 66.67% of people who receive the alerts said they don't read them closely

My Role

  • Designed and analyzed three rounds of surveys on awareness, layout and components, and additional information
  • Compared the current message with five alternatives and proposed a standard wording
  • Designed a web view showing CCTV locations and a possible travel radius with the Kakao Map API
  • Experimented with image generation from descriptions using InsightFace embeddings, IP-Adapter and Stable Diffusion 1.5

Result

  • Alternative 5 drew the most positive responses in the second survey and was chosen as the improved layout
  • The third survey confirmed a preference for a photo-centric UI, leading to a web design with a top photo, a map view, and reporting, saving and sharing
  • Presented the graduation project research at SCI2026 (2026.05), selected as an Outstanding Presentation Paper
Third survey: photo-centric vs map-centric UI
Third survey: photo-centric vs map-centric UI
Web improvement: CCTV locations and possible travel radius
Web improvement: CCTV locations and possible travel radius