Enhanced STE3D-CAP: A novel 3D CAPTCHA family

Abstract

With the growth of the Internet, its wide-ranging services are increasingly being threatened by adverse and malicious attacks. CAPTCHAs have emerged as a standard security countermeasure against Internet attacks such as distributed denial of service attacks and botnets. However, many CAPTCHA schemes themselves have been found to be susceptible to automated attacks. The task of designing a good CAPTCHA scheme is still an open and challenging question, as a good CAPTCHA must fulfil two fundamental requirements; namely, it must be secure against automated attacks whilst being human usable. This paper presents STE3D-CAP-e, a human usable text-based CAPTCHA that is robust against a variety of attacks. STE3D-CAP-e adopts a novel 3D CAPTCHA approach designed to capitalise on the inherent human ability to perceive depth from stereoscopic images. By presenting CAPTCHA challenges using stereoscopic images, humans can distinguish the main text from the background clutter in 3D. The various issues that were considered and addressed in the design of STE3D-CAP-e are described, along with a formal definition of its underlying AI problem family. This paper also presents analysis of STE3D-CAP-e in terms of its security and usability. © 2012 Springer-Verlag.