| |
Click here for full text:
Doodles for Authentication: Recognition and User Study Results
Govindarajulu, Naveen Sundar; Madhvanath, Sriganesh
HPL-2008-36
Keyword(s): Doodles, Authentication, Comparative user study, Dynamic Time Warping, L7 Features
Abstract: Traditional means of computer based authentication based on username and password combinations become unwieldy as the number of password accounts one manages increases. The average computer user needs to remember a large number of text username and password combinations for different applications, which places a large cognitive load on the user. While biometric login based systems can free the user from remembering password information, acceptance of such systems is low due to privacy concerns. We propose the use of personalized hand-drawn "doodles" for authentication. Since doodles can be easier to remember than text passwords, the cognitive load on the user is reduced. Our method involves recognizing doodles by matching them against stored prototypes using handwritten shape matching techniques. To demonstrate the concept we have built a system which uses doodle authentication to login into password protected websites through a web browser. We report accuracy results for our doodle recognition system. We also discuss the results of a user study we conducted to compare the doodle based login system with a text password based login system. We finally conclude with a summary of next steps. Publication Info: Submitted to 2008 International Conference on Intelligent User Interface, Canary Islands, Spain, Jan 13-16, 2008
6 Pages
Back to Index
|
Printable version
