Passdoodles; a Lightweight Authentication Method
Christopher J Varenhorst, Max Van Kleek, Larry Rudolph · 2004
This paper investigates the use of unique finger traces, or doodles, as a means of authentication in a pervasive environment. Velocity here is investigated as means to uniquely identify a doodle. A blurred distribution grid created from combined training samples and the variance across this grid is also used for recognition. These three systems used together have produced accurate results for a population of ten users. The research presented here may have applications to hand writing and drawing recognition as well. 1 Lightweight Authentication Mechanisms New demands of authentication today are simplicity and effortlessness. Biometric technologies offer a partial solution but are be too robust for the relatively small issues such as personalization. Privacy concerns and trust also inhibit public acceptance due to the inextricable ties of biometric imprinting. The remainder of this paper discusses the design and implementation of a lightweight “passdoodle” system where a unique finger trace or doodle is used to quickly identify users in an integrated intelligent (or pervasive) computing environment. 1