The Android Things flower that smiles with you
Smile, and the world smiles with you — or, in this case, a laser-cut flower running Android Things on a Raspberry Pi does.
The aim of the Expression Flower is to “challenge the perception of what robotics can be while exploring the possibility for a whimsical experience that is engaging, natural, and fun.”
Tl;dr: cute interactive flower. No Skynet.
The flower is powered by Google’s IoT platform Android Things, running on a Raspberry Pi, and it has a camera mounted in the centre. It identifies facial expressions using the ML Kit machine learning package, also from Google. The software categorises expressions, and responds with a specific action: smile at the flower, and it will open up its petals with a colourful light show; wink at it, and its petals will close up bashfully.
The build is made of laser-cut and 3D-printed parts, alongside off-the-shelf components. The entire build protocol, including video, parts, and code, is available on hackster.io, so all makers can give Expression Flower a go.
Seriously, this may be the easiest-to-follow tutorial we’ve ever seen. So many videos. So much helpful information. It’s pure perfection!
Machine learning and Android Things
For more Raspberry Pi–based machine learning projects, see:
- Build your own Pokédex, for Pokémon-hunting funtimes
- The Santa detector, for detecting Santa, obviously
- There’s Waldo, for hunting Waldo/Wally in a crowd
And for more Android Things projects, we highly recommend:
- BrailleBox, for displaying the latest news in Braille
- The Android Things Candy Dispenser, for candy…at a price
- The Augmented-reality projection lamp, for, well, it does a lot of stuff, so check it out
Aaaand, for getting started with all things Android on your Raspberry Pi, check out issue 71 of The MagPi!
Raspberry Pi Staff Janina Ander
Julien de la Bruère-Terreault
Great to see different projects using machine learning on the Pi!
Check-out the Rock-Paper-Scissors game I made using python machine learning libraries and running 100% on the Pi (no pre-trained ML model from Google or others): https://github.com/DrGFreeman/rps-cv (shameless plug…)
This is awesome. So awesome that I am currently writing it up for today’s blog post. Thank you!