Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

Make your likes visible on Facebook?

Connect your Facebook account to Prezi and let your likes appear on your timeline.
You can change this under Settings & Account at any time.

No, thanks

An Efficient User Verification System via Mouse Movements

Paper presented by Nan Zheng, Aaron Paloski, Haining Wang from The College of William and Mary
by

Krithika Chandramouli

on 6 February 2014

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of An Efficient User Verification System via Mouse Movements

User authentication and verification are critical
Hijacked online sessions on social networks, online forums can be misused
Password theft
Secret questions- one time authentication

Ideal system: Accurate, Quick in response, Difficult to forge
Behavioral biometrics - 1999

Keystroke dynamics:
detailed timing information that describes exactly when each key was pressed and when it was released as a person is typing at a computer keyboard

Graphical passwords:
user clicks on certain points on screen (an image is shown), and the sequence of clicks is the password
Pros and Cons
Static verification:
about 100 seconds for 20 clicks

Continuous verification:
an average of 1.89 mins for one click to happen => 37.73mins for verification
but 3.03 mins with partial movements included
Previous work
Image by Tom Mooring
An Efficient User Verification System
via Mouse Movements

why angle based metrics?
Platform independence
not affected by the OS, screen size, resolution, mouse
Distance: the accumulated deviation of two distributions over all intervals
Biometric Authentication
centered on "who you are"
a key or card can be lost or stolen
biometrics can neither be lost nor taken away from you
physiological: fingerprint, retina scan-
special hardware
behavioral: keystroke dynamics, mouse dynamics-
use human computer interaction (HCI) get data from common input devices like keyboard, mice etc
Mouse movement and characterization
Data collection through routine computing activities like surfing the net, playing games, chatting etc:
controlled set- 30 users from diverse backgrounds in a normal but controlled environment
field set: ~1000 users on an online forum over 1hr window
tuples
(
action_type
, t, x, y)
mouse move, mouse click

Authentication
Angle Based Metrics
Direction for 2 points AB is the angle between AB and horizontal (x)
Angle of curvature for 3 points A,B,C is the angle between AB and BC (y)
Curvature distance for A,B,C is ratio of length(AC) to perpendicular from B to AC
Unique
similar for same users on different computers, different among different users
distance for a users curvature angles with himself and with others
System Architecture
record user's
mouse
movements
compute angle-based
metrics from
raw data
SVM


threshold and majority
Implemented SVM using LIBSVM, an open sourced tool that can be integrated with Matlab, python,Ruby etc
Threshold
output of SVM above threshold implies a decision of +1 => impostor
an output below threshold => decision -1 => true user
Choosing a majority
half of the samples belong to user, the other half to an 'impostor'
train the users profile 2
n
+1 times, with different sets
the final label is the one with majority
>n
votes
impostors classified as +1 and true users as -1
Evaluation of the system
Profiled on one platform and tested on the same platform
Decision making
Evaluation of the system
False Reject Rate FRR
a true user rejected as an impostor
False Accept Rate FAR
an impostor accepted as a true user
Block of clicks - point and click mouse actions
Verification time
Profiled on one platform and tested on another
innovative concept
well written paper
well thought of - performance evaluation, space consideration
"the verification system will be installed...."
including a snapshot of the UI
better structuring of the paper, maybe

Questions
Nan Zheng, Aaron Paloski, Haining Wang
Dept of Computer Science
The College of William and Mary
Full transcript