BS ISO/IEC 19795-1:2006

Foreword
Introduction
1 Scope
2 Conformance
3 Normative references
4 Terms and definitions
   4.1 Biometric data
   4.2 User interaction with a biometric system
   4.3 Personnel involved in the evaluation
   4.4 Types of evaluation
   4.5 Biometric applications
   4.6 Performance measures
   4.7 Data presentation curves
   4.8 Statistical terms
5 General biometric system
   5.1 Conceptual diagram of general biometric system
   5.2 Conceptual components of a general biometric system
   5.3 Functions of general biometric system
   5.4 Enrolment, verification & identification transactions
   5.5 Performance measure
6 Planning the evaluation
   6.1 General
   6.2 Use of other parts of ISO/IEC 19795
   6.3 Determine information about the system
   6.4 Controlling factors that influence performance
   6.5 Test subject selection
   6.6 Test size
   6.7 Multiple tests
7 Data collection
   7.1 Avoidance of data collection errors
   7.2 Data and details collected
   7.3 Enrolments
   7.4 Genuine transactions
   7.5 Identification transactions of users enrolled in
        the system
   7.6 Impostor transactions
   7.7 Identification transactions of users not enrolled
        in the system
8 Analyses
   8.1 General
   8.2 Fundamental performance metrics
   8.3 Verification system performance metrics
   8.4 (Open-set) Identification system performance metrics
   8.5 Closed-set identification
   8.6 Detection error trade-off/Receiver operating
        characteristic curves
   8.7 Uncertainty of estimates
9 Record keeping
10 Reporting performance results
   10.1 Fundamental metrics
   10.2 Verification system metrics
   10.3 Identification system metrics
   10.4 Closed-set identification system metrics
   10.5 Reporting test details
   10.6 Graphical presentation of results
Annex A (informative) Differences between evaluation types
Annex B (informative) Test size and uncertainty
   B.1 Confidence intervals and test size assuming independent
        identically distributed comparisons
        B.1.1 Rule of 3
        B.1.2 Rule of 30
        B.1.3 Number of comparisons to support a claimed
              error rate
   B.2 Variance of performance measures as a function of
        test size
   B.3 Estimates for variance of performance measures
        B.3.1 General
        B.3.2 Variance of observed false non-match rate
        B.3.3 Variance of observed false match rate
   B.4 Estimating confidence intervals
        B.4.1 General
        B.4.2 Bootstrap estimates of the variance and
              confidence intervals
        B.4.3 Subset sampling
Annex C (informative) Factors influencing performance
   C.1 General
   C.2 List of factors
        C.2.1 Population demographics
        C.2.2 Application
        C.2.3 User physiology
        C.2.4 User behaviour
        C.2.5 User appearance
        C.2.6 Environmental influences
        C.2.7 Sensor and hardware
        C.2.8 User interface
   C.3 Examples for reporting
        C.3.1 Finger position
        C.3.2 Illumination
        C.3.3 Glasses
        C.3.4 Dirt on platen
        C.3.5 Weather
Annex D (informative) Pre-selection
   D.1 Pre-selection algorithm performance
Annex E (informative) Identification performance as a
                      function of database size
Annex F (informative) Algorithms for generating ROC,
                      DET and CMC curves
   F.1 Algorithm for ROC and DET
   F.2 Algorithm for generating CMC
Bibliography

Establishes general principles for testing the performance of biometric systems in terms of error rates and throughput rates for purposes including prediction of performance, comparison of performance, and verifying compliance with specified performance requirements.