Intro to Biometrics

1
Intro to Biometrics

• History
• Why?
• Components

2
Brief History

• Fingerprints on ceramics
• Physical descriptions
• Identifying marks?
• Behavioural descriptions
• Accent/Pronunciation
• Gait
• 19th century Fingerprints
• Modern/automation
• Since 1970s

3
Why?

• Convenience
• Reduced risk of compromise
• Decreased Costs
• Increased Acceptance

4
Biometric Traits

• Genetic
• Inherited features
• Phenotypic
• Embryonic development
• Random
• Behavioural
• Learned/Acquired
• Difficult to vary in adulthood

5
Components

• Data Acquisition
• (Enrollment)
• Transmission Channel
• Data Storage
• Signal Processing
• Decision Policy

6
Data Acquisition

• Trait presented to system
• Digitisation
• (Compression)
• (Parametrisation)
• Encryption

7
Signal Processing

• Extracting relevant features
• Removal of noise
• Extract information carrying features
• (Segmentation)
• Remove unwanted data
• (Normalisation)
• (Quality Assessment)
• Template/Model Creation
• For enrolment
• Output Score

8
Decision Policy

• Verification easier than Identification
• FAR vs FRR
• Can influence choice of biometric
• Crossover Rate
• Equal Error Rate
• Lower is better

9
Enrolment

• User sumits data for template/model creation
• Number of tokens
• Device, User, Environment, Performance
• Quality score
• Threshold setting
• Failure to Enrol Rate (FTER)
• Varies across biometric
• Varies across subjects
• Training?

10
Template Management

• Updates
• Event logging
• Storage
• Local
• Network
• Portable device
• Type affects template management
• Database size and architecture