Transcript handgeometry04

Hand Geometry
BIOM 426
Instructor: Natalia A. Schmid
February 11, 2004
1
Outline
•
•
•
•
•
•
•
•
Motivation
Acquisition systems
Enrollment
Verification
Feature Extraction
Metrics
Applications
Privacy
February 11, 2004
2
References
Not much open literature is available.
Much information is in the form of:
• Patents (for example: Miller’71, Sidlauskas’88)
• Application-oriented descriptions (see IEEE Spectrum no 2, 1994)
• Exclusion: prototype system described by Jain et al. [4]
• Web pages of Recognition Systems and Biomet.ch
• Tutorials (for example, BFC)
February 11, 2004
3
Motivation
Attractive points:
•
•
•
•
Almost all of the working population has hand;
Exception processing can be easily engineered;
Measurements are easily collectable;
Non-intrusive compared to iris or retinal scan;
Simple method of sensing
Computations are easy => system is easy to build
Easy to integrate with other biometrics as fingerprint
Storage efficient (9 bytes)
February 11, 2004
4
Evolution
• First devices (1960s) were
electromechanical. (Miller’s
“Identimation”)
- measures length of 4 fingers
- used in nuclear weapon industry
- was retired in 1987
The existing hand geometry systems
rely on visual images of the hand.
• In the mid-1980’s Sidlauskas
developed electronic 3D profile
identification apparatus.
- capacity 20,000 users
- processing time is 1.2 sec. (1994)
- weight is 4.5 kg (1994)
- 9-byte representation
February 11, 2004
5
Acquisition systems
Features:
- finger length, width, thickness,
curvatures and relative location of
features.
Scanners use:
- CCD camera, infrared LEDs,
mirrors and reflectors.
- No surface details, no color, no
fingerprint lines is recorded.
- Top and side views.
32,000 pixel field
February 11, 2004
6
Acquisition systems
Scanners use:
- Optical path approx. 11 inches
between camera and platen.
Dimensions:
- 8-1/2 inches square by 10 inches in
height.
Scanner takes:
- 96 measurements
Hand scanner optics.
Microprocessor converts:
- 9-byte template
February 11, 2004
7
Enrollment
During enrollment:
- pins (pegs) help user to position
his/her hand
- user places his/her hand 3-5 times
- scanner averages measurements and
stores in the database
Quality of enrollment affects FRR
Factors:
- platen heights
- training (for example, “landing an
airplane scenario”)
Template averaging:
- updating template after user is verified
February 11, 2004
8
Verification
• User types PIN (key pad)
• Places hand on the platen
Scanner
- takes measurements
- extracts features
- compares previous template with
the input template
- generates a similarity score
February 11, 2004
9
Feature Extraction
Typical image: black-and-white
Features: finger length, width,
thickness, curvatures and relative
location of features
February 11, 2004
10
Feature Extraction
An example feature set for hand geometry [4].
February 11, 2004
11
Metrics
• Euclidean distance:
dE 
• Absolute distance:
d
2
(
q

r
)
 i i  E
i 1
d
d A   | qi  ri |  A
i 1
Example:
User 2 = (71, 63, 70, 61, 74, 56, 56, 52, 281, 362, 268, 278, 243, 136)
User 2 = (69, 63, 74, 62, 73, 57, 57, 55, 276, 366, 259, 282, 245, 141)
User 15 = (55, 56, 63, 53, 60, 47, 48, 47, 249, 303, 258, 268, 241, 152)
d E (User2, User2 )  14.1421
d A (User2, User2)  42
d E (User2, User15)  75.8222
d A (User2, User15)  203
February 11, 2004
12
Market
Access Control
Used to access Health clubs, Day care centers, Laboratories, Prisons, etc.
Time & Attendance
Application ranges from coal mines to clean rooms.
Personal Identification
Newark and Toronto airports; Food Services systems at the University of
Georgia
(See more on http://www.recogsys.com/)
February 13, 2004
13
Applications
• 70,000 HandReaders are installed throughout the world.
• The 1996 Olympic Games used HandReaders to protect access to Olympic Village
(65,000 people were enrolled; 1 million transactions were handled over 28 days).
• Since 1991, at San Francisco Airport, HandReaders produced more than 100 million
verifications (180 doors and 18,000 employees).
• In the United Kingdom, Her majesty’s Prisons rely on the HandReaders for prisoner
and visitor tracing.
• Colleges (ex. University of Georgia) use HandReaders for on-campus meal programs,
safeguard access to dormitories and protect their computer centers.
• Over 20,000 Owens Illinois employees punch in and out each day using the
HandReader.
• Krispy Cream Doughnuts uses HandReaders for tracking employee hours at over 30
individual stores.
February 13, 2004
14
1. Privacy Issues
Hand geometry is used to verify identity.
Templates cannot be “reverse engineered” to identify users.
2. Operation by Disabled People
Hand scanners can be used for scanning left hand (palm up).
Could be enabled for blind persons to use.
February 13, 2004
15
Positives and Negatives
(See [5] pp. 146-147).
February 13, 2004
16
Available Databases
1. University of Bologna database
http://bias.csr.unibo.it/research/biolab/bio_tree.html
2. MSU hand geometry database.
3. Ongoing project at WVU (multi-modal biometrics)
February 13, 2004
17
References
1. Biometrics: Personal Identification in Networked Society, A. Jain et al. Edt.
2. Hand: give me five by D. Sidlauskas in “Vital signs of identity,” IEEE
Specrtum, February 1994, pp.24 - 25.
3. D. P. Sidlauskas, “3D hand profile identification apparatus,” US Patent No.
4736203, 1988.
4. A. K. Jain, A. Ross, and Sh. Pankanti, “A Prototype Hand Geometry-based
Verification System,” Proc. of 2nd Int’l Conf. on Audio- and Video-based
Biometric Person Authentication, Washington D.C., pp. 166-171, March 22-24,
1999.
5. R. M. Bolle, et al., Guide to Biometrics, Springer, New York, 2004, pp. 45-47.
6. http://www.recogsys.com/
7. http://www.biomet.ch/ (two-finger verification)
February 13, 2004
18
Preprocessing
February 13, 2004
19
Feature Extraction
February 13, 2004
20
Feature Extraction
Matlab Code:
>> IM = imread(‘filename’,‘tiff’); % read tiff-file
>> BW = im2bw(IM,0.75);
% binarization
>> size(IM)
% provides info. about image size
>> mask = zeros(512,640);
% creates image filled with zeros
>> mask (260,190:450) = 1;
% fills line with ones
>> Feature = (1-BW).*mask;
% extracts feature
>> length(find(Feature > 0))
% finds feature length in pixels
February 13, 2004
21