[ad_1]
With the long summer evenings comes the perfect opportunity to dust off your old boxes of circuits and wires and start building something. If you lack inspiration, you might be interested in the practical guide of the artist and engineer Dan Macnish on building a Doodle camera powered by AI with a thermal printer, Raspberry pi, a pinch of Python and the Google Quick Draw dataset. 19659002] "Playing a day with neural networks for object recognition, I wondered if I could push the concept of Polaroid further and ask the camera to reinterpret the image by printing a drawing instead of a faithful photograph Macnish wrote on his blog about the project, called Draw This.
To do this work, Macnish tapped into Google's object recognition neural network and in the data set created for the Google Quick game, Draw! By linking the two systems with python code, Macnish was able to recognize by creating real images and print the best matching doodle in the Quick, Draw ! dataset
But since the output doodles are limited to the data set, there may be a discrepancy between what the camera "sees" and what it generates for the photo.
"You point and shoot; the best camera interpretation of what he's seen, "writes Macnish. "The result is always a surprise." A healthy food selfie could turn into a huge hot dog. "
If you want to get started, Macnish has downloaded the instructions and code needed to build this project on GitHub.
Source link
