Thursday, September 17, 2009

Google acquires ReCaptcha as book-scanning aid | Relevant Results - CNET News

Seeing a pattern?…

The idea behind a captcha is to confuse a computer, but computers are also confused by some words written in fonts used long ago. ReCaptcha offers two words, one of which is a captcha it already knows, and one of which is a word it doesn't know. The thinking is that if you get the first word right, you're likely a human and you're also probably going to get the second one right.

It can then pool all the answers for the second word and declare with a reasonable amount of certainty that the second word is what most people think it is, thereby updating the vocabulary of participating book scanners. This is of obvious interest to Google, currently bent on scanning as many books as it can find.

Google acquires ReCaptcha as book-scanning aid | Relevant Results - CNET News

No comments: