FriendLens

Face Detection and Tracking

Inspiration

You're in a crowded room, full of unfamiliar faces. If only you could quickly scan the room to see if any of your friends are there.... With FriendLens, you can!

How many times have you found yourself at a party or a networking session or any event, really, wondering (1) if anyone you know is there and (2) what is that guy's name? You met him through a friend of a friend, and you're friends on Facebook now, but....

What it does

FriendLens allows you to hold up your phone to a crowd and find where your friends are!

How we built it

FriendLens is an iOS application that uses the Microsoft Cognitive Face API to detect and identify the people the camera sees in real-time. The back-end is hosted on a Django server.

Challenges we ran into

Trying to integrate the various parts was the biggest challenge, as well as working with APIs and tools that none of us had experience in before.

Face-tracking is incredibly non-trivial, and we tried many different approaches, from built-in iOS functionalities to OpenCV to a hybrid approach, which we settled on.

Accomplishments that we're proud of

This is a great step toward making the world more connected by allowing people to find each other in-person from their interactions online.

What we learned

Lots of things :o

What's next for Friend Finder

In addition to identifying friends, we want to be able to identify friends of friends. We would also like to be able to send friend requests to friends of friends directly from the app, but this would require integration with Facebook app APIs.