On predictive models and user-drawn graphical passwords

摘要: In commonplace text-based password schemes, users typically choose passwords that are easy to recall, exhibit patterns, and thus vulnerable brute-force dictionary attacks. This leads us ask whether other types of (e.g., graphical) also attack because tending memorable passwords. We suggest a method predict model number such classes for systems where created solely from user's memory. hypothesize these define weak subspaces suitable an dictionary. For user-drawn graphical passwords, we apply this with cognitive studies on visual recall. These motivate set complexity factors reflective symmetry stroke count), which classes. To better understand the size and, thus, how they might be, use “Draw-A-Secret” (DAS) scheme Jermyn et al. [1999] as example. analyze DAS under convenient parameter choices show can be combined apparently popular have bit sizes ranging 31 41—a surprisingly small proportion full space (58 bits). Our results quantitatively support suggestions employ measures, rules or guidelines proactive checking.