DooDB: A Graphical Password Database Containing Doodles and Pseudo-Signatures

Abstract

Touch screen-enabled devices are proliferating in the communications and entertainment markets. In this scenario, the use of graphical passwords for user validation is receiving an increasing interest in the last years. Unlike in other fields of research on automatic user authentication, such as biometrics, there are no public databases of graphical passwords usable for research purposes (to the extent of our knowledge). In the present work, the recently captured DooDB database is introduced. This database comprises two sub corpora: doodles and simplified signatures (pseudo-signatures). These data were produced by 100 users, who were asked to draw with their fingertips over a mobile device touch screen. Forgeries are also included in the database. A quantitative analysis of both datasets is first performed. Preliminary verification experiments using the two kinds of graphical passwords are reported.