Visual Passwords


January 28, 2019

Graphical passwords are an authentication mechanism for computer systems. The difference between a VP and the currently dominant alphanumeric password is that with a VP, a user’s password is represented by where that user clicks on an image. Thus, an application using graphical passwords for authentication would show a picture to the user. The user would then click in a number of places on the picture, and the coordinates of the clicks would be stored by an application. During authentication, the user has to click on the established points. (The system, of course, allows for configurable error tolerance, since it is not realistic to expect a person to click on exactly the same point each time.)

Graphical passwords attempt to deal with the same problem as do the usual alphanumeric passwords. However, graphical passwords also try to address some drawbacks that are inherent in alphanumeric passwords–hopefully without introducing any drawbacks of their own.

There are some basic requirements that are built into authentication systems based on “what you know.” The password should be easy to remember by the legitimate user, but should be hard to guess by everybody else. Unfortunately, those requirements are in conflict. If the password is easy for a user to remember, very likely it is made up of some word and/or some significant number for that user. The word can be either some significant name of a person or place, which can either be found in a dictionary or from basic knowledge about the person, and the same can be said about dates. To summarize, alphanumeric passwords are generally easy to guess. Also, harder passwords or the ones for many different systems are usually written on stickie notes, which makes them less secure.

A graphical password offers a much larger key-space than an alphanumeric one, which is limited to roughly 64 ASCII characters. For example, if we have a 600-by-800 image and an error tolerance of 10 pixels, it would result in 4,800 possibilities. Also, the graphical password is much harder to write down or even tell to some other person. Last but not least, another benefit of graphical passwords is the cued-recall, which helps users to remember a password based on the picture displayed, and not just on memory alone.

From a usability point of view, we conduct experiments to see whether graphical passwords are at least as easy for people to use as alphanumeric passwords. We address both the technical security, which involves the transmission and storage of the password in a secure manner, as well as the user security, which involves an analysis of whether people use the system in a secure or insecure way. The latter is an analysis of how people choose the graphical passwords, and whether they are vulnerable to guessing or dictionary attack.

Participants: Nasir MemonAleksandr Brodskiy

Resources:Graphical Password Homepage


Nasir Memon Professor: Computer Science and Engineering New York University Tandon School of Engineering 10.095, 2 MetroTech Center, 10th Floor Brooklyn, NY 11201

Tel: 718-260-3970
Fax: 718-260-3609
Google Voice: 1-480-2MYPROF‏
Skype: (212) 203-0608‏; Username: evilproffy
Email: memon@nyu.edu