Post Syndicated from Alex Bate original https://www.raspberrypi.org/blog/facepunch-facial-recognition/
How it works
dekuNukem uses a Raspberry Pi 3, the Raspberry Pi camera module, and an OLED screen for the build. You don’t strictly need to include the OLED board, but it definitely adds to the overall effect, letting you view your daily and weekly screen time at a glance without having to access your Raspberry Pi for data.
As dekuNukem explains in the GitHub repo for the build, they used a perf board to mount the screen and attached it to the Raspberry Pi. This is a nice, simple means of pulling the whole project together without loose wires or the need for a modified case.
This face_recognition library lets the Pi + camera register your face. You’ll also need a well lit 400×400 photograph of yourself to act as a reference for the library. From there, a few commands should get you started.
Uses for facial recognition
You could simply use facepunch for its intended purpose, but here at Pi Towers we’ve been discussing further uses for the build. We’re all guilty of sitting for too long at our desks, so why not incorporate a “get up and walk around” notification? How about a flashing LED that tells you to “drink some water”? You could even go a little deeper (though possibly a little Big Brother) and set up an “I’m back at my desk” notification on Slack, to let your colleagues know you’re available.
What would you do with facial recognition on a Raspberry Pi?