Using A Two Dimensional Colorized Barcode Solution for Authentication in Pervasive Computing

1
Using A Two Dimensional Colorized Barcode
Solution for Authentication in Pervasive
Computing

• William Claycomb and Dongwan Shin
• New Mexico Institute of Mining and Technolgoy
• June 28, 2006

2
Acknowledgement

• New Mexico Tech Graduate Student Association

3
Agenda

• Introduction and Background
• Motivation
• UbiColor
• Testing and Results
• Analysis
• Practical Applications and Advantages
• Future Work
• Questions

4
Introduction

• The use of mobile devices in pervasive computing
  environments is growing
• The types of interactions available are also
  increasing
• The security of these interactions needs to be
  addressed
• Particularly in one-to-one situations

5
One-to-one Communications
Alice
Bob
?
6
Background

• Authentication and Key Exchange
• Devices need to authenticate for key exchange to
  occur
• Resurrecting Duckling
• Location limited channels

7
Background

• Using Digital Cameras
• Visual tags
• Capture an image of a tagand decode
  informationcontained within
• Establishing a communication channel
• Demonstrative identification

8
Background

• Using Visual Tags for Authentication and Key
  Exchange
• UbiCode
• Developed by Shin and Im, 2005
• Two Dimensional barcode for ad-hoc device
  authentication

9
UbiCode
10
Motivation to Improve

• Ubicode limitations
• Small pixel size
• IPv4 address space
• 128 bit hash algorithm
• Mobile device improvements
• Higher resolution cameras
• Bright, colorful displays

11
Improvement UbiColor

• Expand barcode capacity by adding color
• Increase pixel size by reducing the number of
  pixels needed to convey information
• Increase the amount of information contained in
  the barcode
• Utilizes existing capabilities of mobile devices

12
UbiColor Color Chart
13
UbiColor

• A UbiColor Visual Tag

14
UbiColor Tag Detail
15
Ubicolor Tag Detail
16
Establishing a Secure Connection
17
Agenda

• Introduction and Background
• Motivation to Improve
• UbiColor
• Testing and Results
• Analysis
• Practical Applications and Advantages
• Future Work
• Questions

18
Testing

• Control Images
• Brightness
• Rotation
• Tag Size within image
• Distortion
• Blurriness and Noise

19
Testing Digital Photos

• Variables
• Room lighting
• Background surface color
• Picture resolution
• Device brightness
• Size of tag within image
• Display device
• Camera

20
Testing Platform and Setup

• Three testing platforms
• Dell Latitude D600
• Dell Axim X50v PDA
• HP iPAQ rx3715 PDA
• Setup
• Three images taken of each configuration
• All images tested on each device
• 576 images 1584 total tests

21
Results

• Display a visual tag

22
Processing Control Images

• Brightness
• Positive brightness 10, 20, 30, 40
• Negative brightness -10, -20, -30
• Rotation
• Positive rotation angle 1, 2, 5, 10, 20, 30
• Negative rotation angle 1, 2, 5, 10
• Size of Visual Tag Within the Image
• Visual tag width to image width 30, 40, 50,
  60, 70, 80, 90
• Distortion
• Horizontal Distortion
• Positive (right side) distortion 5, 10, 15,
  20, 30, 40, 45
• Negative (left side) distortion 5, 10, 15,
  20, 30, 40, 45
• Vertical Distortion
• Positive (bottom) distortion 5, 10, 15, 20,
  30, 40, 45
• Negative (top) distortion 5, 10, 15, 20,
  30, 40, 45

23
Processing Digital Photos - Time
24
Performance Data Extraction
25
Performance Establishing Secure Communication

• Algorithm time lt 1 second
• Network time 1 6 seconds average

?
26
Agenda

• Introduction and Background
• Motivation to Improve
• UbiColor
• Testing and Results
• Analysis
• Practical Applications and Advantages
• Future Work
• Questions

27
Analysis

• Cameras make a big difference
• Current state of digital camera technology in
  smart devices
• 5.0 megapixels coming soon!

28
Analysis Control Images

• Rotation Issues

29
Analysis Digital Photos

• Resolution
• Higher resolution picks up more detail
• This is not a good thing

30
Analysis Digital Photos

• Brightness
• Half-brightness better overall, but not that much
  difference

31
Analysis Digital Photos

• Background room lighting
• Device background color
• Neither make that big of a difference

32
Analysis Digital Photos

• Display Device
• Major flaw discovered in implementation
• Color detection algorithm uses first match,
  rather than best match.

33
Analysis Digital Photos

• Tag size within image
• Better chance of error with large tag

34
Agenda

• Introduction and Background
• Motivation to Improve
• UbiColor
• Testing and Results
• Analysis
• Practical Applications and Advantages
• Future Work
• Questions

35
Practical Application Scenario

• ATM
• Capture visual tag image from ATM
• Establish secure connection
• Authenticate user
• Several storage methods for this
• Interact with ATM via mobile device
• ATM responds when appropriate

36
Practical Application Scenario

• ATM (cont)
• Advantages
• Less physical use of service device
• More options for physically locating the device
• Personalization is possible
• Additional physical security
• No need for ATM card
• Protection from skimming

37
Practical Application Scenario

• Vending Machine
• Capture image of visual tag from vending machine
• Establish secure connection
• Interact with vending machine
• Receive product information
• Make purchases using digital cash, credit card,
  store credit, etc.

38
Practical Application Scenario

• Vending Machine
• Advantages
• Product selection can be externally controlled
• Payment information can be stored on the device,
  but that is not required
• Electronic receipts

39
Advantages of this Approach

• This approach provides a secure communication
  channel
• Particularly useful for sensitive transactions
• Easily adaptable to various networking
  environments
• Adaptable to various computing platforms
• Does not require presence of CA
• External connection not required, unless further
  user authentication is necessary
• This approach does not require the use of device
  discovery protocols

40
Advantages of this Approach

• Security
• Man-in-the-middle attack prevented
• Misidentification of tag information is not a
  threat
• The incorrectly identified host will not have the
  correct public key
• Keys change with every instance of communication
• The of pre-printed, superimposed visual tags for
  and attack is possible, but not likely

41
Agenda

• Introduction and Background
• Motivation to Improve
• UbiColor
• Testing and Results
• Analysis
• Practical Applications and Advantages
• Future Work
• Questions

42
Future Work

• Design Changes
• Corners
• Visual Tag Identification
• Alignment
• Error correction
• Exists in other 2Dbarcodes already

43
Additional Colors?

• Additional Colors
• Adding 8 colors wouldincrease the bitrate by
  33
• Can the applicationdistinguish thesecolors?

44
Future Work

• Application Changes
• Improve tag identification
• Improve color recognition
• Two-way communication

45
Future Work

• Group Communication

46
Questions