News 

Captcha (Completely Automated Public Turing test to tell Computers and Humans Apart) is the commonly used method for preventing spam bots and other pernicious software tools from automating registering on websites. But researchers at Carnegie Mellon University realised that the technology could also be used to get humans to decipher old texts that were proving beyond the capabilities of text recognition software.

The reCaptcha system works by breaking the text down into captcha-sized chunks and distributing

ADVERTISEMENT

them to participating websites, of which there are already about 40,000. When a website user completes a captcha form, the entered text is sent back to the reCaptcha database. Each snippet is checked a few times to ensure veracity before a final version is stored.

Despite some problems with short words and a lackadaisical approach when it comes to matching capitalisation and punctuation, the system is already more than 99% accurate.

It also has one major benefit for participating websites. Unlike standard, computer-generated captchas, the reCaptchas have proved completely impenetrable to the new captcha-reading tools that spammers have begun deploying. This is because the reCaptchas are generated from real type and consequently have variations, unevenness, irregularities that are not present in the standard captcha.

For more information — and to help out by completing a few reCaptcha’s — go to recaptcha.net/learnmore.html.