Continuous authentication
Jump to navigation
Jump to search
Continuous authentication
Als je daar in jan 2016 op zoekt, krijg je van Google de volgende suggesties.
Laten we zo de paragrafen indelen. Of we ons lucky moeten feelen is nog maar de vraag.
Definitie en overview (veel studiemateriaal gelinkt)
Contents
Mobile continuous authentication
Continuous authentication using biometric keystrokes
Noorse technische verhandeling over expirimenten uit 2009
Een deel van de conclusies
- The results in the previous section are promising, in the sense that an intruder will be, on average, locked out fairly quickly. Obviously this average should be as low as possible to make the system more secure.
- One way of improving thesecurity would be to look at the typing behaviour for different applications. It is not hard to see that informal chatting on media like MSN, Yahoo or Skype will have a completely different typing behaviour compared to programming in C++, entering data in MS Excel, preparing slides for presentations or using a web browser. In our experiment we already collected information on the used application, so a next stepin the analysis will be to factor the application into the penalty and reward function.
- One more thing that needs to be mentioned here is that, in our experiment, there were absolutely no restrictions on the behaviour of the participants. This also included their freedom of choice of the hardware to use, so some users could have been using a normal keyboard, while others where using a laptop, where the typing behaviour will probably be very different. Also the layout of the keyboard (US, NO, or any other) might have been different for different participants in the experiment.
- This might imply that if an intruder will try to use the computer of a victim, his typing behaviour might change from his own normal typing behaviour because he has to adapt to the hardware and layout of the keyboard of the victim. It is not clear beforehand if such a change of typing behaviour would work in favor of the intruder or against him. This again needs to be investigated in future research.
- Finally in the future, we might improve the performance of the continuous authentication system, by combining keystroke dynamics with mouse usage. This can be done at two levels. First we can measure the particular statistics of the typing rhythm and the mouse usage and use this kind of information for authentication. This is more or less the fusion of a continuous keystroke dynamics and a continuous mouse dynamics system. Second, we can also measure the interaction between
keyboard and mouse, meaning that we model the user behaviour at the computer and use that for authentication.
