Breaking an animated CAPTCHA scheme

  • Authors:
  • Vu Duc Nguyen;Yang-Wai Chow;Willy Susilo

  • Affiliations:
  • Centre for Computer and Information Security Research, University of Wollongong, Australia;Advanced Multimedia Research Laboratory, School of Computer Science and Software Engineering, University of Wollongong, Australia;Centre for Computer and Information Security Research, University of Wollongong, Australia

  • Venue:
  • ACNS'12 Proceedings of the 10th international conference on Applied Cryptography and Network Security
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

CAPTCHAs have become a ubiquitous security countermeasure to protect online web-services against automated attacks. However, attackers have managed to successfully break many existing CAPTCHA schemes. Animated CAPTCHA schemes have been proposed as a method of producing CAPTCHAs that are more human usable and more secure. The addition of the time dimension is supposed to increase the robustness of animated CAPTCHAs. This paper investigates the robustness of HelloCaptcha, an animated text-based CAPTCHA scheme with a total of 84 different variations. In this paper, we show that simple techniques can be used to extract important information from the animation frames of an animated CAPTCHA. Our approach essentially reduces the animated CAPTCHA into a traditional single image CAPTCHA challenge. Furthermore, the methods presented in this paper can be generalized to break other animated CAPTCHAs.