Draw This, the camera that uses a neural network to convert a photo into a cartoon

Artificial intelligence, neural networks and artificial vision are terms that are gaining more and more ground and that technology is advancing even faster than a decade ago. years. But, this is not surprising because, we've seen how companies are striving to create more and more standalone devices, and cameras with greater precision and that with the help of these technologies give surprising results, so far we had not known camera based on a neural network

Currently, most cameras are designed to capture scenes in the best possible way, but that will now make another sense, since Dan Macnish, has built an instant camera called Draw This, which uses a network of neurons to convert pictures into cartoons.

Macnish took advantage of data from Google Quick Draw, an online game that uses a neural network to try to guess what people were trying to draw, and generated a database. more than 50 million drawings of adhesive figures to feed your camera.

Macnish's camera, works on a Raspberry Pi, with a thermal printer and other electronic devices that help him work in the best Polaroid style. However, this camera does not have a viewer or preview screen.

Curiously, if you take a selfie, this camera may scribble a bicycle wheel, or if you take a picture of your dog It could become a spaceship, as described by its creator:

One of Fun things about this reinvented polaroid, that is you can never see the original image. You aim and shoot, and a caricature appears; the best camera interpretation of what he's seen. The result is always a surprise. A food selfie with a healthy salad could become a huge hot dog, or a photo with friends could be changed to goat.

As you can see, these resulting images reflect how the neural network interprets the scene in front of it and although it does not. projecting what you see in the form of a caricature seems to us a very important starting point for us to reach in the not too distant future.