A new algorithm for graphical user authentication based on rotation and resizing

Abstract

It is now beyond any doubt that USER AUTHENTICATION is the most critical element in the field of Information Security. To date, Text Based Password Authentication (TBPA) has shown some difficulties that users have tended to write passwords down manually or save them on hard disc. This tendency is caused by the passwords being strong and thus difficult to memorize in most cases. This has inadvertently given rise to security issues pertaining to attack. Graphical User Authentication (GUA) has two symbiotic pillars as its foundation: USABILITY & SECURITY. The macro-concept of GUA is based on the human psychological factor that is images are more readily committed to memory than would TBPA’s. All GUA algorithms are made up of both usability and security aspects. Unfortunately none of the existing algorithms are able to cover both of these aspects at the same time. The main purpose of this thesis is an algorithm that combines the usability & security features. by Rotating and Resizing the Images. This algorithm process on the images which user select during login phase and show a different picture the original password of the user into a GUA. Whereas previously the GUA’s were at the log-in stage, this algorithm would be at the registration stage. On the other hand, unlike previous graphical user authentication algorithms, in this algorithm the images selected as passwords in the registration phase will be different from the images in log-in phase. The usability features of implemented algorithm will be checked by a questionnaire survey based on ISO usability features that are defined in literature review. The attack resistance features will be evaluated by calculating the “Password Entropy” and “Password Space” and compared with the previous algorithms. These are the two major evaluation methods for graphical passwords. A comparison table for analysis will be created to compare previous algorithms with the proposed algorithm.