User Password Using Random Art

Abstract

Although research in security has made great progress over the past few years, most security systems still suffer by failing the account human factors. Humans are slow and unreliable at processing and comparing long meaningless password string; and people have limitation difficulties in remembering secure passwords or Personal Identification Number (PINs). In this paper, fundamentals weaknesses of knowledgebased authentication schemes are addressed, and how the usability and security of the user authentication systems can be improved by replacing text string with structured images are investigated. The prototype solution of this replacement utilizes Hash Visualization function (HVF), which uses Random Art (RA). Random art was development by Andrej Baure, and is based on an idea of genetic art by Michael Witbrock and John Mount. The basic idea is to use a binary string s as a seed for a random number generator. The randomness is used to construct a random expression which describes a function generating the image mapping each image pixel to a color value. RA has been constructer to satisfy all the properties of the HVF, which are image generation, ease of computation, near-one-way property, regularity property, complexity property, and entropy property. This approach relies on recognition-based authentication rather than recall-based authentication, since it authenticates a user through recognizing the previously seen image. We examine the requirement of recognition-based system and proposed the Random Image Password System (RIPS). The RIPS is relatively more reliable and easier to use than the traditional recall-based schemes. Furthermore, it has the advantage that it prevents users from choosing weak