Finder_

Opt-in web application for private cameras to help identify and locate missing persons and wanted individuals using facial comparison.

Comment

Inspiration

When brainstorming ideas for hackUMBC, Finder_ was one of the first we came up with and could not let go of. HackUMBC has provided a great reason for us to challenge ourselves with new skills and hopefully create something that can help locate missing and kidnapped individuals in the future.

What it does

Finder_ uses your camera and starts a recording where every few seconds, a frame is taken from the video feed. This frame is then compared to the profile images of our missing person's database using Face++. If a match 75% or better is made, the database is updated with the captured frame, the location of the camera, and time of capture. The admin access is also alerted of this match, allowing administrators to confirm if the match is valid and send the information to necessary parties.

How we built it

We started with a schema for the entire project, breaking it down into multiple parts. The home, camera and admin pages, the facial comparison backend, and the database.

Challenges we ran into

Originally, we planned to also use AWS in order to communicate between our MySQL database and our site backend, but AWS's JavaScript was depreciated, preventing us from being able to use it. So, we had to find alternative ways to serve up the required images. Instead after many failed workarounds, we used Google Cloud Buckets to send the images to Face++ as links to avoid 414 URI error (too much information).

Accomplishments that we're proud of

-Learning new libraries in a short amount of time -Executing a design plan and ending up 24 hours later with one almost identical to where we started -Pushing ourselves to our limits and working hard

What we learned

We learned about fast track project development and the numerous challenges that come with it. If one person doesn't carry their weight, the entire project can fall apart. UMBC's Hackathon gave us the opportunity as a team to push ourselves and each other not only with what we know, but how we work in a team.

What's next for Finder_

We plan on connecting Finder_ with multiple social media sites by scraping public photos. Those photos would then be subjected to the same facial comparison framework we already have in play in hopes of finding matches to missing, kidnapped, and wanted individuals.

Built With

Share this project:

Updates