With lengthy summer season evenings comes the right alternative to mud off your previous bins of circuits and wires and begin to construct one thing. If you’re brief on inspiration, you is likely to be enthusiastic about artist and engineer Dan Macnish’s how-to information on constructing an AI-powered doodle digicam utilizing a thermal printer, Raspberry pi, a splash of Python and Google’s Quick Draw information set.
“Playing with neural networks for object recognition one day, I wondered if I could take the concept of a Polaroid one step further, and ask the camera to re-interpret the image, printing out a cartoon instead of a faithful photograph.” Macnish wrote on his blog concerning the undertaking, referred to as Draw This.
To make this work, Macnish drew on Google’s object recognition neural community and the data set created for the sport Google Quick, Draw! Tying the 2 techniques along with some python code, Macnish was capable of have his creation acknowledge actual photographs and print out one of the best corresponding doodle within the Quick, Draw! information set
But since output doodles are restricted to the info set, there may be some discrepancy between what the digicam “sees” and what it generates for the photograph.
“You point and shoot – and out pops a cartoon; the camera’s best interpretation of what it saw,” Macnish writes. “The result is always a surprise. A food selfie of a healthy salad might turn into an enormous hot dog.”
If you wish to give this a go for your self, Macnish has uploaded the directions and code wanted to construct this undertaking on GitHub.