Feeds

Google boffins beat own Captchas

The StreetView numbers game

Reducing security risks from open source software

A group of Google scientists working on extracting numbers from StreetView images has discovered that their technology can also match humans at solving captchas.

The aim, according to their research paper (at Arxiv, here), was to automatically extract accurate street number data from StreetView images so as to improve Google Maps location information.

Prior work, the researchers write, had worked on extracting individual numbers from an image, identifying each number, and then reassembling the whole street number. This, however, is inefficient, so the group let by Ian Goodfellow focussed on taking an entire image and identifying all the numbers in it.

Testing their model on Google's StreetView House Numbers dataset (which contains 200,000 numbers), the researchers found they were able to match human accuracy of 98 per cent with “95.64 per cent coverage”.

To achieve that accuracy, the researchers spent six days training Google's DistBelief neural network modelling framework. That training was then applied to all the house numbers Google holds – well into the tens of millions. The less constrained dataset reduced coverage down to 89 per cent while holding accuracy at the 98 per cent “equal to a human” threshold.

Google's Street View number recognition

Getting numbers out of images is easy, says Google

The same model was then tested against Google's reCAPTCHA puzzle, achieving 99.8 per cent accuracy. The researchers write that while this doesn't render Captchas useless, “the utility of distorted text as a reverse turing test by itself is significantly diminished”.

As Google's Vinay Shet writes in this blog post, “the act of typing in the answer to a distorted image should not be the only factor when it comes to determining a human versus a machine”, and Google itself is reducing its “dependence on text distortions as the main differentiator between human and machine,” using Captchas instead to “perform advanced risk analysis”. ®

Mobile application security vulnerability report

More from The Register

next story
LibreSSL RNG bug fix: What's all the forking fuss about, ask devs
Blow to bit-spitter 'tis but a flesh wound, claim team
Microsoft: You NEED bad passwords and should re-use them a lot
Dirty QWERTY a perfect P@ssword1 for garbage websites
Manic malware Mayhem spreads through Linux, FreeBSD web servers
And how Google could cripple infection rate in a second
NUDE SNAPS AGENCY: NSA bods love 'showing off your saucy selfies'
Swapping other people's sexts is a fringe benefit, says Snowden
Own a Cisco modem or wireless gateway? It might be owned by someone else, too
Remote code exec in HTTP server hands kit to bad guys
British data cops: We need greater powers and more money
You want data butt kicking, we need bigger boots - ICO
Crooks fling banking Trojan at Japanese smut site fans
Wait - they're doing online banking with an unpatched Windows PC?
NIST told to grow a pair and kick NSA to the curb
Lrn2crypto, oversight panel tells US govt's algorithm bods
prev story

Whitepapers

Top three mobile application threats
Prevent sensitive data leakage over insecure channels or stolen mobile devices.
The Essential Guide to IT Transformation
ServiceNow discusses three IT transformations that can help CIO's automate IT services to transform IT and the enterprise.
Mobile application security vulnerability report
The alarming realities regarding the sheer number of applications vulnerable to attack, and the most common and easily addressable vulnerability errors.
How modern custom applications can spur business growth
Learn how to create, deploy and manage custom applications without consuming or expanding the need for scarce, expensive IT resources.
Consolidation: the foundation for IT and business transformation
In this whitepaper learn how effective consolidation of IT and business resources can enable multiple, meaningful business benefits.