Each year in China alone, 20,000 to 200,000 children went missing, kidnapped or trafficked. They are usually forced to labor, auctioned illegally for adoption, or even entrapped in gangs. People witness them on the street, but the reporting and identification process is frustratingly time-consuming. Convenient and timely identification check is essential to further actions. We aspire to seek solutions in the light of face recognition technology.
What it does
Our web-based app allows users to match their cloud photo repositories with a missing person database. Upon confident matches, the app will prompt identifications to the user for further investigation. In detail, the app does:
- Get photo access autherization from repositories (Google photos for the submission)
- Identify distinct faces in each photo
- Match each face with identities in the databases
- Return photo pairs of 2 similar faces (and mark them)
- Render paired results on the website and allow users to decide if we have a match
How we built it
Challenges we ran into
- Comparing different Face API and work around their conditions, which has been a lot
- Node.js - a new web tool for the entire team.
Accomplishments that we're proud of
We are proud to have come up with this powerful idea and made a preliminary product within such short time with new tools.
What we learned
All components of a web application on node; in particular we learned to write async programs. App development in team.
What's next for Find Missing People
- More repositories: Facebook, iCloud, or video searching/ cameras?
- Not as an app but as an extension? In fact, we see the app to be 'hidden' behind the scene and only message the user or relevant authorities when a match happens. It might a recursive program on the photo storage platforms.
- Better people database. The application is not limited to missing people, but also criminals at large.
- Distributed systems and map-reduce for faster calculations