When the history books are written after the robot uprising, should we survive, a sad chapter will be devoted to the day humans were bested at yet another simple feat by one of our silicon overlords; finding Waldo.
According to The Verge, creative agency Redpepper has created There’s Waldo – robotic metal arm made by Ufactory with a Vision Camera Kit and a floppy prosthetic hand attached to it, all powered by a tiny Raspberry Pi computer.
The camera takes a photo of the page, which then uses OpenCV to find the possible Waldo faces in the photo. The faces are then sent to be analyzed by Google’s AutoML Vision service, which has been trained on photos of Waldo. If the robot determines a match with 95 percent confidence or higher, it’ll point to all the Waldos it can find on the page. –The Verge
Google’s Cloud AutoML, available since January, allows users to train their own AI tools without prior coding knowledge using a drag-and-drop tool for image recognition purposes. The tool can be trained for a variety of cases, such as categorizing photos.
Redpepper creative technologist Matt Reed told The Verge via email: “I got all of the Waldo training images from Google Image Search; 62 distinct Waldo heads and 45 Waldo heads plus body. I thought that wouldn’t be enough data to build a strong model but it gives surprisingly good predictions on Waldos that weren’t in the original training set.”
Reed was inspired by Amazon Rekognition’s ability to recognize celebrities, and wanted to experiment on a similar system which supported cartoons. He had no prior experience with AutoML, and it took him about a week to code the robot in Python. –The Verge
What will robots take the fun out of next?
Read on ZH