Protection through Intelligent and Multimedia Captchas

The Complete Automatic Public Turing test to tell Computers and Humans Apart (CAPTCHA) is a modern implementation of the Turing test. A CAPTCHA asks a series of questions which are supposed to be relatively easy for mature human beings but hard for machines to answer. A CAPTCHA, as its name suggests, is generally used as a tool to distinguish human beings from machines when the corresponding party is invisible in a remote site. There are many CAPTCHAs proposed and widely used in many Internet applications, with varied strength and performance. In this article, a review of current CAPTCHAs is presented. Analysis and comparisons are provided for various CAPTCHAs. Two types of new CAPTCHAs, intelligent and multimedia CAPTCHAs, are proposed. Our experiments have shown that the newly proposed CAPTCHAs have stronger protection than most existing CAPTCHAs and they can be applied to various web applications, such as e-commerce. DOI: 10.4018/jaras.2012040103 International Journal of Adaptive, Resilient and Autonomic Systems, 3(2), 44-58, April-June 2012 45 Copyright © 2012, IGI Global. Copying or distributing in print or electronic forms without written permission of IGI Global is prohibited. it is now understood how to engineer Scatter Type challenges to meet legibility and difficulty targets (Baird, Moll, & Wang, 2005). It is known that humans perform better in reading familiar text than unfamiliar text (Pope & Kaur, 2005). Also, there is a large gap in the ability of reading and interpretation between humans and machines. This allows designers to design some CAPTCHAs by exploiting tradeoffs between the familiarity of challenge strings and degradation of the image. In this paper, the current CAPTCHAs are categorized and reviewed, and their strength and features are summarized. Based on the existing concepts, two new CAPTCHAs: intelligent and 3-D multimedia CAPTCHAs, are introduced and prototyped. They will be shown to have stronger strengths of protection than that of the existing CAPTCHAs. Their application to e-commerce will also be discussed. 2. CURRENT RESEARCH ON CAPTCHAS CAPTCHAs can be classified into two main categories. The first operates on visual information, i.e., using images and the second one is audioinformation based. Visual-based CAPTCHAs are the most widely applied and they counter attack three machine intelligences i.e., Optical Character Recognition (OCR), visual pattern recognition and a combination of both (Pope & Kaur, 2005; Szeliski, Avidan, & Anandan, 2000). The list below summarizes the most popular CAPTCHAs. A brief introduction on each of the CAPCHAs is presented. • Text based ◦ Gimpy ◦ EZ-Gimpy ◦ PessimalPrint ◦ Baffle Text ◦ Scatter Type • Graphic based ◦ Bongo ◦ Pix • Audio based • Mouse intervention CAPTCHA 2.1. Visual CAPTCHAs