Point a camera at an object, click the shutter button, and you can save an image. Some cameras do this better than others. And some can print your images right away, Polaroid-style.
And then there’s Draw This. It’s DIY camera that uses artificial intelligence to identify the item you’re photographing and print a line drawing of a matching object.
Take a picture of a cat and it’ll spit out a sheet of paper with a drawing of a cat. Photograph a flower and you’ll get a simple flower drawing. And snap a photo of a person and you’ll get a crappy cartoon version of a person.
While Draw This is probably a pretty lousy way to save memories, it’s a pretty impressive project that combines some off-the-shelf technologies in surprising ways.
Developed by Dan Macnish, consists of a camera uses a Raspberry Pi with a camera and a 16GB SD card, a thermal printer, and a cardboard case. It then uses a Google image recognition model to identify objects and Google’s quickdraw dataset to translate object categories into drawings.
If you’re interested in building your own Draw This camera, the instructions and code are available at github.
Or if you don’t have time to build your own camera, you can still waste some time playing with some of the underlying technology by checking out Google’s Quick Draw web app which attempts to recognize your crappy drawings.
via Hackaday
This has great potential for an enterprising artist to take lots of photographs of objects (people, cats, vases, cut flowers, cans of soup etc), print out the drawings, then pin them on all the visible sides of display blocks, mount those on a large base, surrounded by waist high posts with cord ropes, get a very big arts council grant, and display them in a modern art exhibit with a suitably erudite and culturally profound title.
Call it WSAPBATSO. Also known as “Why Starving Artists Produce Better Art Than Subsidized Ones.”
could be a cool time waster on a wet winters day 🙂