h a l f b a k e r yLoading tagline ....
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Write a neural net type program to analyse mouse movements, much like hand writing. You could sit down and trace out your signature with the mouse pointer and it would recognise you and log you in - no more passwords to remember.
No new hardware to buy and almost impossible to fake.
Authentication via mouse signature.
http://www.ananova....enu=news.technology
Kinda Baked. [silverstormer, Oct 04 2004]
[link]
|
| |
Doing the same thing with your keyboard timing is quite common, it is refered to as "keystroke dynamics". I found a few folks mentioning mouse dynamics as an intrusion detection technique on the net, but nobody seems to have baked it yet. |
|
| |
How do you know that mouse movements are unique? |
|
| |
I have seen this done before. The closest I can find link wise...(linky) |
|
| |
I need some clarifcation before I can decide what to think about this... Are you suggesting that people actually sign their name using the mouse? Or is your idea to use analyze a person's (possibly) unique mouse dynamics when they are using the computer? |
|
| |
Could be defeated by a PS2/USB device that replays a mouse signature that you've previously recorded. Probably less secure than a password, which you can at least change on a regular basis. |
|
| |
Well theres ways around most security procedures - maybe couple it with a password. Not necessarily sign their name, could be a shape or something. It could just monitor your mouse movements continuously whilst you're logged on and detect whether someone else is using your session, you could disable it at login if you have a sprained wrist or something. I'm sure that deviation from start point to finish point, average curve deflection during mouse movements and acceleration graph statistics would be pretty unique user to user. |
|
| |
My bank used to require random mouse movements together with the usual passwords at login, but more to show you're a human bean trying to authenticate rather than some bot. I've also seen it used to generate an encryption key (when setting up a Blackberry PDA). |
|
| |
// Could be defeated by a PS2/USB device that replays a mouse signature that you've previously recorded. // |
|
| |
You could use a PS2/USB device to log keystrokes to steal passwords too. Whuddya gonna do... |
|
| |
We’re sorry Dave, but our computer has determined that you are not human enough to proceed. |
|
| |
[tpoo22], should bean = being? |
|
| |
So this means that I'm the only one who moves the mouse pointer wherever on the screen I put my sight? |
|
| |
no pericles, it's the meandering path you take to get there that identifies you (in theory). |
|
| |
I think it's actually about your specific habits, peculiarities, and nuances of how you move and use the mouse being characterized and recorded and subsequently used to identify you as the current user of the mouse. |
|
| |
(I hadn't given it much thought before, but I rarely look directly at the spot that I intend to move the mouse pointer to unless I'm doing some sort of graphics work.) |
|
| |
Can be done but the program will depend upon the type of task done by the user as well as the experience. I recently observed people change their mouse movements once they are experts in a certain task. Also different ypes of mouse movement exist. Mostly a initial ballistic quick movement and then a slow precise one when target selection is considered. Searching is typically related to the nature of search (list, group) and the intent. Intra user statistics are pretty unique as far as this ballistic part is considered (some overshoot some undershoot). Difficult to guess all possible type of movements the user might do. An initial qualitaitve description might be a good idea for quantifying the unique characteristics. |
|
| |