Remember those “Where’s Waldo” books? Well, now there’s a robot designed to find him in just about any picture. It’s called “There’s Waldo,” and it consists of a Raspberry Pi-controlled uArm Swift Pro equipped with a Vision Camera Kit for facial recognition. First, the camera snaps a photo of the page, uses OpenCV to find the possible Waldo faces in the photo, and then the faces are then sent to be analyzed by Google’s AutoML Vision service trained on photos of Waldo. If the 95 percent confidence match or higher is confirmed, it points out all the Waldos it can find on the page. Continue reading for more pictures and information.
“Google’s Cloud AutoML has been available since January to let users train their own AI tools without any previous coding knowledge. The drag-and-drop tool lets anyone create an image recognition tool, which has a variety of use cases such as categorizing photos of ramen by the shops they came from. You can catch a glimpse of this process in the video above, in which different photos of Waldos are fed into the software,” reports The Verge.