There's Waldo is a robot that finds Waldo

Follow by Email
We built a little robot called "There's Waldo" to test the capabilities of Google's new AutoML Vision service. We've found that technologies can be unapproachable, and irrelevant by extension, to many people—so we learn ahead of the curve, and show our work in fun ways, to demonstrate what's possible. There's Waldo is a robot built to find Waldo and point at him. The robot arm is controlled by a Raspberry Pi using the PYUARM Python library for the UARM Metal. Once initialized the arm is instructed to extend and take a photo of the canvas below. It then uses OpenCV to find and extract faces from the photo. The faces are sent to the Google Auto ML Vision service which compares each one against the trained Waldo model. If a confident match of 95% (0.95) or higher is found the robot arm is instructed to extend to the coordinates of the matching face and point at it. If there are multiple Waldos in a photo it will point to each one. While only a prototype, the fastest There's Waldo has pointed out a match has been 4.45 seconds which is better than most 5 year olds. Want to learn more about what we do? Visit:

Comments from Youtube

Adlin Ling : Find John Connor next.

explosu : Do you know how many 5 year olds are going to be out of a job because of you?!?!?!?!

Alex Mecca : Can you make one to help find my dad.

René Bigum : Now, imagine if this was used in mass surveillance. Oh wait...

Nolan St. Martin : Let's see it try the last page of the Great Waldo Search, ie. the Land of Waldos.

Ario : Meme potential here


Andrey Kitsen : About time used technology for something useful.

Guille Pérez : Sorry for the silly question but in the last picture, why did the arm point to the little Waldo while it also has a bigger picture of his face on the letter? Shouldn't the model have identified this picture as well?

XxRapid DenvetxX : Back from daily dose of internet (DDI)

Saturn Mike : It couldn't handle the harder pages.

Ryan Carty : This feels like something Michael Reeves would make

Meeperz : 0:45 *boi*

blaaka33 : Amazing, how many hours did it take to build this , how many images did you use to train it, is it much simpler than doing real photographs/people compared to this or isn't it that far off?

wilkiezoo : For it's next trick, Jimmy Hoffa

Eon : I'm more impressed with how accurate it is at making the pointing finger land on waldo.

avi bank : Can't stop watching. Probably going to do this as part of my postgrad research. Will ref you guys!

I Eye : I could use a robot like that to find my key XD

Syj Barbados : Incredibly amazing.

AyeYerMa : Hmm, would be better to see the entire machine functioning instead of close up shots of a rubber hand slapping the page. How do I know some random dude isn't manually just moving the arm?

258th Hi Guy : I can't believe they set up a waldo to find Waldo

Ian Kim : Jeez, not even Waldo is safe from the robot uprising.

Jimmy Punch Lines : Waldo on sweat !

Paulo Gurgel Carlos da Silva : Só perguntam: ONDE está Wally? Ninguém nunca pergunta: COMO está Wally?

The Jollybox : Poor Waldo, he keeps being slapped by the robot :(

Dustin Benzel : Someone was salty they couldn't find Waldo, so they created a robot that could find Waldo for them.

Matthew McDonald : Is there anything they won't build a robot to replace humans for?! We're not safe!

james pruett : Inventor should give a shout out to Intel for open-sourcing OPENCV which I imagine they are using... or to the HARR algorithm.

jokerjjg : Another step towards Skynet. Thanks.

T H : But can he find Flight MH 370 ?

Joaquín Fernández Gálvez : Man!! if you built a robot with a wood base (OMG) please paint the base so we can't see that is done with wood. 4.45 seconds is fastest one?? the vision systems that I use to install take 250 ms to analyze 10 images and found 6 patterns in each one....but keep on it (and try open GL)

Alexander Riccio : What's the breakdown of those 4.45 seconds?

Jucab Nubster : Daily Dose of Internet sent me! You've made it!

Craig Tyle : Overhead, without any fuss, the stars were going out.

Balfour : Greatest invention ever

BikeStuff : its little hand is so cute.

LillyBillyDilly _/ _/ : But the real question is: Can it see John Cena? 🤔

Zapdog : Train it to find my dignity.

crimescene25 : can finally find my peenis!

Nate Barlow Jr. : I wish like hell I would have had the source code for this 2 years ago.

Daniël Boon : help send code

ecafretni resu : awesome demo video!

BlakeRPhillips : My kids could do it quicker. They have memorized the locations. It's great fun to compete in a "who found each item" the quickest game... Gotta find those scrolls, bones, books, etc.!

MedEighty : And, at 0:50, it didn't notice the bigger Waldo in the postage stamp. Is this a fake video or was the software configured to pick only pictures of a specific size?

roverbudy : amazing

Lone Star Rocker : Just like the human brain!

marshwiggle360 : HAHAHAHA

Shelley Wu : Funny it’s programmed to use middle finger instead of index finger to do the pointing

Mats Kr : I guess it's better than most 5 year olds :)