Recaptcha Does Double Duty
Chances are good that if you're online alot, the little red box and marked through words at left are already familiar to you. It's a system called Captcha (Completely Automated Public Turing test to tell Computers and Humans Apart). First of all, you can see why they shorten it to Captcha. Second of all, this little service is employed by many websites for a very good reason. You, the person who wants to leave a comment, or otherwise use a website, are forced to parse out the obscured and distorted words, thereby proving that you're a human, and not a spambot looking to harvest addresses and send out junk mail.
Good on you, Captcha. The folks who run the project say that it's used around 100 million times a day. But as the infamous advertisement for Ginsu Knives goes, "But wait! There's more!"
According to a report in the journal Science, the Recaptcha folks, part of the Computer Science Department at Carnegie Mellon University, are going a step further. There's a new scheme that allows websites to choose words that have been deemed unreadable by the optical scanning software being used to digitize old texts. You see, damaged texts or faded ink can mean that up to 20 percent of a given text is unreadable by digital means. So, those words are farmed out in the Captcha system -- the same word to more than just one person of course -- and the results are compiled in a database.
Think of it as crowd-sourcing for the archivally-minded.
The results are impressive. Better than 99 percent accuracy, according to Recaptcha. So far, 40,000 websites have signed up for the project, and more than 440 million words have been resolved.
And that means more texts available online for our general perusal and edification. Delightful. Actually, it brings a tear to the eye, seeing humans and computers being able to work together like this.
Eat that, er...read that, Skynet.
![Reblog this post [with Zemanta]](http://img.zemanta.com/reblog_e.png?x-id=defd3bca-3987-4da2-abfb-83d8c0581b1c)
Comments