How Google’s reCAPTCHA System Is Advancing Self-Driving Cars
If you go on the internet, you’ve probably had to solve a CAPTCHA or reCAPTCHA security test at some point or another. Google acquired the system in 2009, and since then, there have been some changes to the program. Have you noticed it likes to ask about cars, buses, crosswalks, and other vehicle-related images lately? Google has not confirmed that it uses the reCAPTCHA system for autonomous vehicles, but here are a few reasons why I think that could be the case. Self-driving cars from Waymo and other brands are improving every day, but the process requires a lot of critical technology and data to improve continuously.
Google allegedly uses reCAPTCHA data for self-driving cars
According to an old Google Security Blog, using reCAPTCHA and Street View to make locations on Maps more accurate was happening way back in 2014. Google says CAPTCHA stands for Completely Automated Public Turing test to tell Computers and Humans Apart. When the CAPTCHA system was updated to reCAPTCHA, it would ask users to find the street numbers found on Google Street View and confirm the numbers matched. Previously, it would use distorted text or letters. Using this data, Google could correlate the numbers with addresses and help pinpoint the location on Google Maps.
A scientific study from Cornell University looked at the technology as well. The paper, titled Multi-digit Number Recognition from Street View Imagery using Deep Convolutional Neural Networks, found that the systems could find and read the house numbers presented on Street View with about 90% accuracy. I remember these being hard to do initially, but the images improved over time.
Medium reports that more than 60 million CAPTCHAs are being solved every day, which saves around 160,000 human hours of work. If these were helping locate addresses, why not also help identify other objects? Help differentiate a bus from a car and even choose a crosswalk over a light pole. This helps image processing take a step in the right direction.
The reCAPTCHA system has changed a lot under Google
Waymo, which is formally known as the Google self-driving car project, calls itself the World’s Most Experienced Driver. It is the world’s first autonomous ride-hailing service, and how wouldn’t this benefit from our bus-versus-crosswalk image selection? This isn’t to say the only program being utilized is the reCAPTCHA system. Perhaps it is the third or fourth layer of checking for errors or confirming something.
We know Google is on top of its AI game and has many dedicated teams for all of these systems. It would only make sense that the reCAPTCHA would also help further the autonomous driving situation. The car can continue on its journey by detecting items in front of the vehicle and differentiating one object from another.
While Google has not confirmed this, maybe it will one day. The next time you fill out a reCAPTCHA, take note of what it is. Are you selecting all of the photos of bicycles, but there are none? Do you have to choose the crosswalks out of nine images showing palm trees, fields, and random streets?
Google and Waymo are still full speed ahead on autonomous driving
As autonomous driving continues to improve, perhaps the reCAPTCHA will move on to another project. Maybe it will just continue to ask more complicated questions like selecting the back of a truck over the back of an SUV. If that helps self-driving systems become better and more accurate in real-life situations, every little bit helps.
What started as a way to keep spammers and scammers at bay has become a helpful technology. Google Maps has improved by leaps and bounds in the last few years, and directions are more precise (save for the guy who followed it off a bridge). It will be interesting to see if the reCAPTCHA puzzles continue to change and evolve as technology does in 2024 and beyond.