Document

Report
BIOMETRICS:
EAR RECOGNITION
Samantha L. Allen
Dr. Damon L. Woodard
July 31, 2012
OUTLINE
I. Biometrics: What Is It?
II. Why Biometrics?
III.Ear Biometrics
IV.How A Biometric System Works
V. Conclusion
WHAT IS IT?
Biometrics
• The science and technology of measuring and analyzing
biological data
• Measures and analyzes human body characteristics for
authentication
• Physical or behavioral characteristics
• Identity access management and access control
BEHAVIORAL CHARACTERISTICS




Keystroke
Voice patterns
Gait
Signature
PHYSICAL CHARACTERISTICS






DNA
Fingerprints
Eye retinas and irises
Facial patterns
Hand measurements
Ear geometry
BIOMETRIC SYSTEM
COMPONENTS
Sensor
Feature
Extraction
Matcher
DATABASE
BIOMETRIC SYSTEM OPERATION
Verification
Identification
• Identity Claimed
• One-to-one Comparison
• Authentication is either
approved or denied.
• No identity claimed
• One-to-many comparison
• Identity is determined
(OR)
• User not being enrolled leads
to fail of identification.
WHY BIOMETRICS
• Biometrics is a method of *direct* human identification as opposed
to identifying humans by their possession of keys or remembering
passwords.
• Preferred method of identification because ID’s and cards can easily
be stolen and passwords are likely to be forgotten or shared.
• Discourages fraud
• Enhances security
DISADVANTAGES TO BIOMETRICS
 Privacy Concerns
 Irrevocable
 Functional Creep
 Output is “matching score” instead of yes/no
BIOMETRIC SELECTION PROCESS
 Permanence
 Performance
 Acceptability
 Distinctiveness
 Circumvention
 Collectability
 Universality
EAR BIOMETRICS BACKGROUND
• Dates back to the 1980’s
• Shape and features of ear
 Unique
 Invariant with age
• Disadvantages
 Affected by occlusions, hair,
and ear piercings
EXAMPLES OF BAD IMAGES
2D VS. 3D EAR BIOMETRICS
• Performance is greatly
affected by pose variation
and imaging conditions
• Images contain less
information
• Contains surface shape
information related to
anatomical structure
• Relatively insensitive to
illumination
• Slightly higher performance
EAR BIOMETRICS APPROACHES
• Approaches
 Global: Whole ear
 Local: Sections of ear
 Geometric: Measurements
HOW A BIOMETRIC SYSTEM
WORKS
• Has this applicant been here before?
• Is this the person that he/she claims to be?
• Should this individual be given access to our system?
• Are the rendered services being accessed by a legitimate user?
HOW A BIOMETRIC SYSTEM
WORKS (CONT.)
HOW A BIOMETRIC SYSTEM
WORKS (CONT.)
• Identifying features of individual are enrolled into system.
• During feature extraction, the application is used to identify
specific points of data as match points
• Match points in database are processed using an algorithm
that translates the information into numeric values or
feature vectors.
• Feature set is compared against the template set in the
system database.
EAR RECOGNITION
DETECTION PROCESS
• Human ear detection is a crucial task of a human ear
recognition system because its performance significantly
affects the overall quality of the system.
 template matching based detection
 ear shape model based detection
 fusion of color and range images and global-to-local registration based
detection
PERFORMANCE METRICS
The following are used as performance metrics for biometric systems:
• False accept rate or false match rate (FAR or FMR)
 Measures the percent of invalid inputs which are incorrectly accepted.
 Probability that the system incorrectly matches the input pattern to a nonmatching template in the database.
• False reject rate or false non-match rate (FRR or FNMR)
 Measures the percent of valid inputs which are incorrectly rejected.
 Probability that the system fails to detect a match between the input
pattern and a matching template in the database.
SUMMER RESEARCH
• Research included exploration of ear recognition
implementation in Matlab.
• 100 pre-processed images, 17 subjects
SUMMER RESEARCH
• Enroll images into database
with different classes for
each person
• Perform ear recognition or
1:1 verification
CONCLUSION
• Ear recognition is still a relatively new area in biometrics
research.
• Potential to be used in real-world applications to
identify/authenticate humans by their ears.
• Can be used in both the low and high security applications and
in combination with other biometrics such as face.
REFERENCES
•
•
•
•
•
•
•
•
D. Hurley, B Arbab-Zavar, and M. Nixon, The Ear as a Biometric, In A. Jain, P.
Flynn, and A. Ross, Handbook of Biometrics, Chapter 7, Springer US, 131150, 2007.
A. Jain, A. Ross, and S. Prabhakar. An Introduction to Biometric Recognition. In IEE
Trans. On Circuits and Systems for Video Technology, Jan. 2004.
R. N. Tobias, A Survey of Ear as a Biometric: Methods, Applications, and Databases
for Ear Recognition.
Carreira-Perpiñán, M. Á. (1995): Compression neural networks for feature extraction:
Application to human recognition from ear images (in Spanish). MSc thesis, Faculty of
Informatics, Technical University of Madrid, Spain.
http://www.advancedsourcecode.com/earrecognition.asp
http://vislab.ucr.edu/PUBLICATIONS/pubs/Chapters/2009/3D%20Ear%2
0Biometrics09.pdf
http://www.security.iitk.ac.in/contents/publications/more/ear.pdf
http://www.technovelgy.com/ct/Technology-Article.asp?ArtNum=98

similar documents