Authenticating humans to computers remains a notable weak point in computer security despite decades of effort. Although the security research community has explored dozens of proposals for replacing or strengthening passwords, they appear likely to remain entrenched as the standard mechanism of human-computer authentication on the Internet for years to come. Even in the optimistic scenario of eliminating passwords from most of today's authentication protocols using trusted hardware devices or trusted servers to perform federated authentication, passwords will persist as a means of 'last-mile' authentication between humans and these trusted single sign-on deputies. This dissertation studies the difficulty of guessing human-chosen secrets, introducing a sound mathematical framework modeling human choice as a skewed probability distribution. We introduce a new metric, alpha-guesswork, which accurately models the resistance of a distribution against all possible guessing attacks. We also study the statistical challenges of estimating this metric using empirical data sets which can be modeled as a large random sample from the underlying probability distribution. This framework is then used to evaluate several representative data sets from the most important categories of human-chosen secrets to provide reliable estimates of security against guessing attacks. This includes collecting the largest-ever corpus of user-chosen passwords, with nearly 70 million, the largest list of human names ever assembled for research, the largest data sets of real answers to personal knowledge questions and the first data published about human choice of banking PINs. This data provides reliable numbers for designing security systems and highlights universal limitations of human-chosen secrets.

Authenticating users by means of passwords is still the dominant form of authentication despite its recognised weaknesses. To solve this, authenticating users with images or pictures (i.e. graphical passwords) is proposed as one possible alternative as it is claimed that pictures are easy to remember, easy to use and has considerable security. Reviewing literature from the last twenty years found that few graphical password schemes have successfully been applied as the primary user authentication mechanism, with many studies reporting that their proposed scheme was better than their predecessors and they normally compared their scheme with the traditional password-based. In addition, opportunities for further research in areas such as image selection, image storage and retrieval, memorability (i.e. the user’s ability to remember passwords), predictability, applicability to multiple platforms, as well as users’ familiarity are still widely possible. Motivated by the above findings and hoping to reduce the aforementioned issues, this thesis reports upon a series of graphical password studies by comparing existing methods, developing a novel alternative scheme, and introducing guidance for users before they start selecting their password. Specifically, two studies comparing graphical password methods were conducted with the specific aims to evaluate users’ familiarity and perception towards graphical methods and to examine the performance of graphical methods in the web environment. To investigate the feasibility of combining two graphical methods, a novel graphical method known as EGAS (Enhanced Graphical Authentication System) was developed and tested in terms of its ease of use, ideal secret combination, ideal login strategies, effect of using smaller tolerances (i.e. areas where the click is still accepted) as well as users’ familiarity. In addition, graphical password guidelines (GPG) were introduced and deployed within the EGAS prototype, in order to evaluate their potential to assist users in creating appropriate password choices. From these studies, the thesis provides an alternative classification for graphical password methods by looking at the users’ tasks when authenticating into the system; namely click-based, choice-based, draw-based and hybrid. Findings from comparative studies revealed that although a number of participants stated that they were aware of the existence of graphical passwords, they actually had little understanding of the methods involved. Moreover, the methods of selecting a series of images (i.e. choice-based) and clicking on the image (i.e. click-based) are actually possible to be used for web-based authentication due to both of them reporting complementary results. With respect to EGAS, the studies have shown that combining two graphical methods is possible and does not introduce negative effects upon the resulting usability. User familiarity with the EGAS software prototype was also improved as they used the software for periods of time, with improvement shown in login time, accuracy and login failures. With the above findings, the research proposes that users’ familiarity is one of the key elements in deploying any graphical method, and appropriate HCI guidelines should be considered and employed during development of the scheme. Additionally, employing the guidelines within the graphical method and not treating them as a separate entity in user authentication is also recommended. Other than that, elements such as reducing predictability, testing with multiple usage scenarios and platforms, as well as flexibility with respect to tolerance should be the focus for future research.