CAPTCHAs: Humans vs. Bots

Abstract

A completely automated public turing test to tell computers and humans apart (CAPT-CHA) offers a way for Web service providers to make some conclusions about whether a "user" is human or robot. Process of CAPTCHA recognition is a combination of efforts, approaches, and software that attempts to increase accuracy to an acceptable level. Of course, it's hard to define this level, but we've found that it often starts from 5 percent accuracy. We've covered the most popular CAPTCHA approaches here but not all of them. Most users know how to use the more commonly used graphical CAPTCHAs, so designers must have serious reasons to migrate to other, more difficult-to-break, approaches. After all, CAPTCHAs aim to determine legitimate users, not alienate them.