Text based graphical password system to obscure shoulder surfing

摘要

Traditional text based password schemes are subjected to dictionary attacks on a very large scale. As a solution, graphical password schemes are a promising alternative to text-based recognition schemes where instead of text, images are chosen for a password. However, these schemes are again affected due to shoulder surfing and less effective due to large dictionary space. This paper focuses on providing a solution based on text-based graphical password techniques, a combination of Déjà vu and Moveable Frame scheme. Déjà vu is further replaced by text-based images than purely graphics. Our research has two main objectives: 1) To thwart shoulder surfing 2) To minimize the search time of pass images on the login screen. The project was implemented as a resulting android application using Android Studio. Results show the proposed scheme of graphical password system with multiple alphabet images is more recognizable and has less cognitive load on user as compared to image based graphical password schemes. The incorrect login rate for multiple alphabet based images with movable frame was 15% more than graphical password schemes thus a better solution to protect shoulder surfing.