Do you need help finding Waldo? Since 4.45 seconds is all it takes for Redpepper’s robot to find him, it has been given the appropriate name ‘There’s Waldo’. What used to be a challenging game for children has now also become a subject to Artificial Intelligence.
The little metal robotic arm is equipped with a Vision Camera Kit that allows for facial recognition. After the camera has taken a photo of the page, Open CV is used to find possible Waldo faces. These faces are then sent to the Google’s AutoML Vision service, and if the robot determines a match with a confidence level of 95% or higher it will point to all the Waldo’s it can find on that page.
However, the remarkable thing is that the creative technologist at Redpepper who oversaw the project is not that trained in coding at all. Google’s AutoML Vision is designed to let users train their own AI tool without having to write a single line of code. For now this service is limited to image recognition, where users can upload a set of images, sit back and let the machine do the work.
There’s Waldo was created to show in a simple way what machine vision can do and to show the customers of Redpepper the services that the company offers. But the same technology could be used in many other cases, like identifying how many pairs of Adidas shoes or Tommy Hilfiger sweaters are being worn at the mall or counting how many people drinking a beer are seen in a photo from a festival. Many people find these ‘tricks’ amusing, others take it more seriously. Some critics take it a little step further, claiming There’s Waldo shows (in a much simpler version though) how systems like Facebook and government surveillance facial recognition software can spot people in a crowd. However, that would be a lot easier for them if you wear the same red-striped shirt every day.
References:
https://www.theverge.com/2018/1/17/16901126/google-cloud-ai-services-automl
