Cyberstudy
PDF p.86 In progress

Biometric Authentication

Open PDF at p.86 22 flashcards

Summary

PDF p.86

Biometric authentication uses unique physiological or behavioral characteristics to verify identity. The process involves enrollment, where a biometric sample is captured and converted into a template, and subsequent authentication, where new scans are compared to the stored template. Key metrics for evaluating biometric systems include False Rejection Rate (FRR), False Acceptance Rate (FAR), and Crossover Error Rate (CER).

In plain words

Supplementary — not from your PDF

Biometrics use your body or behaviour, such as fingerprint or face. The system stores a template and compares new scans against it. Tuning is a balance: too strict rejects real users (FRR), too loose accepts impostors (FAR). The crossover point (CER) measures overall quality.

Detailed explanation

PDF p.86
  • Enrollment Process
    • Sensor Module: Acquires the biometric sample from the user.
    • Feature Extraction Module: Creates a mathematical template representing unique identifiers.
    • Authentication: User is re-scanned, and the new scan is compared to the template. Access is granted if they match within a defined tolerance.
  • Evaluation Metrics
    • False Rejection Rate (FRR)
      • Definition: Percentage of legitimate users not recognized (Type I error).
      • Impact: Causes inconvenience to users.
    • False Acceptance Rate (FAR)
      • Definition: Percentage of interlopers accepted (Type II error).
      • Impact: Can lead to security breaches.
    • Crossover Error Rate (CER)
      • Definition: Point where FRR and FAR are equal.
      • Importance: Lower CER indicates more efficient and reliable technology.
  • Performance Factors
    • Throughput (Speed): Time required for template creation and authentication, crucial for high-traffic areas.
    • Failure to Enroll Rate (FER): Incidents where a template cannot be created during enrollment.
    • Cost/Implementation: Varies by scanner type; some are expensive or difficult to use on mobile devices.
    • User Acceptance: Concerns about privacy, intrusiveness, and accessibility for those with disabilities.
  • Common Biometric Methods
    • Fingerprint Recognition
      • Technology: Uses capacitive cells or optical cameras to detect unique ridge patterns.
      • Advantages: Inexpensive, nonintrusive, straightforward.
      • Challenges: Moisture or dirt can affect readings.
    • Facial Recognition
      • Technology: Records indicators like eye distance and nose width using optical and infrared cameras.
      • Advantages: Defeats spoofing attempts with photos.
      • Challenges: Accuracy can be affected by lighting and angles.

Important terms

taken from the text above
Sensor Module
Acquires the biometric sample from the user.
Feature Extraction Module
Creates a mathematical template representing unique identifiers.
Authentication
User is re-scanned, and the new scan is compared to the template. Access is granted if they match within a defined tolerance.
False Rejection Rate (FRR)
Percentage of legitimate users not recognized (Type I error).
False Acceptance Rate (FAR)
Percentage of interlopers accepted (Type II error).
Crossover Error Rate (CER)
Point where FRR and FAR are equal.
Throughput (Speed)
Time required for template creation and authentication, crucial for high-traffic areas.
Failure to Enroll Rate (FER)
Incidents where a template cannot be created during enrollment.
Cost/Implementation
Varies by scanner type; some are expensive or difficult to use on mobile devices.
User Acceptance
Concerns about privacy, intrusiveness, and accessibility for those with disabilities.
Technology
Uses capacitive cells or optical cameras to detect unique ridge patterns.
FRR False Rejection Rate FAR False Acceptance Rate CER Crossover Error Rate FER Failure to Enroll Rate

Examples & real-world scenarios

Supplementary — not from your PDF
  • Fingerprint readers on laptops.
  • Face recognition with infrared to resist photos.
  • A lower CER meaning a more accurate system.

Scenario

A data center door scanner is tuned to let almost nobody in by mistake. Staff are now often rejected and queue at the door. FAR went down but FRR went up, which is the trade-off.

Common mistakes

Supplementary — not from your PDF
  • Mixing up FAR (impostor accepted, a security risk) and FRR (genuine user rejected, an inconvenience).
  • Forgetting privacy and acceptance concerns with biometrics.

Practical skills

Supplementary — not from your PDF
  • Explain FAR, FRR and CER with an example.

What I should remember

Key Points PDF p.86
  • Enrollment Process
    • Sensor Module: Captures biometric sample.
    • Feature Extraction: Creates unique template.
    • Authentication: Compares new scan to template.
  • Evaluation Metrics
    • FRR: Legitimate users not recognized.
    • FAR: Interlopers accepted.
    • CER: Balance point of FRR and FAR.
  • Performance Factors
    • Throughput: Speed of template creation and authentication.
    • FER: Failure to create a template.
    • Cost/Implementation: Expense and ease of use.
    • User Acceptance: Privacy and accessibility concerns.
  • Common Methods
    • Fingerprint Recognition: Inexpensive, nonintrusive.
    • Facial Recognition: Accurate, resists spoofing.