5 FpVTE 2012—Class B

Results in this section involve variations of Identification Flat captures (i.e., right slap, left slap, and thumb slaps, otherwise known as finger positions 13–15 or a 4-4-2 configuration). Probes of right slap, left slap, right and left slap, and a complete Identification Flat were searched against an enrollment database of 3 000 000 subjects containing all ten fingers in an Identification Flat configuration.

The datasets in this section are equivalent to those used in NIST FpVTE 2012 (Class B). Detailed information about FpVTE 2012 can be found in NIST IR 8034.

Notes:

  • No examiner extended feature set data was provided with the images.
  • Slap segmentation, if required, was performed by hisign+0015.

5.1 Template Generation

The approximate total number of records that underwent template generation along with a tally of records that failed to process are shown in Table 5.1. Each template was generated by a single function call providing hisign+0015 all of the listed image types.

Table 5.1: Summary of template generation for FpVTE 2012—Class B.
Image Contents Template Type Failure to Extract ≈Total
Left Slap Probe 4 30 000
Right Slap 6 30 000
Left + Right Slap 10 30 000
Identification Flats 553 30 000
Reference 46 068 3 000 000

5.2 Search

The probe templates from Table 5.1 were searched against an enrollment database of 3 000 000 subjects containing images as specified in the reference template row of Table 5.1. Approximately one-third of the probes had a corresponding mate in the enrollment database.

5.2.1 DET

The DET plot in Figure 5.1 show the tradeoff of errors of hisign+0015 when searching each probe set from FpVTE 2012—Class B against enrollment database of 3 000 000 subjects where, for approximately one-third of the probes, a single mated identity consisting of all ten fingers in an Identification Flat configuration was present. Tabular versions of FNIR at select FPIR can be viewed in Table 5.2.

DET when searching probe templates from FpVTE 2012---Class B against an enrollment database of Identification Flats.

Figure 5.1: DET when searching probe templates from FpVTE 2012—Class B against an enrollment database of Identification Flats.

Table 5.2: FNIR values from the DET plotted in Figure 5.1.
Probe Content FPIR  ≤ 0.001 FPIR  ≤ 0.005 FPIR  ≤ 0.01
Left Slap 0.0218 0.0212 0.0209
Right Slap 0.0209 0.0207 0.0207
Left + Right Slap 0.0195 0.0195 0.0194
Identification Flats 0.0360 0.0360 0.0360
Table 5.3: Similarity score thresholds from the DET plotted in Figure 5.1.
Probe Content FPIR  ≤ 0.001 FPIR  ≤ 0.005 FPIR  ≤ 0.01
Left Slap 744.0 721.0 712.00
Right Slap 739.0 720.0 711.00
Left + Right Slap 829.5 800.5 784.00
Identification Flats 833.0 801.0 786.25

5.2.2 CMC

The CMC plot in Figure 5.2 show the FNIR of hisign+0015 when searching each probe set from FpVTE 2012—Class B against enrollment database of 3 000 000 subjects where, for approximately one-third of the probes, a single mated identity consisting of all ten fingers in an Identification Flat configuration was present. Tabular versions of FNIR at select ranks can be viewed in Table 5.4.

CMC when searching probe templates from FpVTE 2012---Class B against an enrollment database of Identification Flats.

Figure 5.2: CMC when searching probe templates from FpVTE 2012—Class B against an enrollment database of Identification Flats.

Table 5.4: FNIR values from CMC plotted in Figure 5.2.
Probe Content Rank 1 Rank  ≤ 2 Rank  ≤ 5 Rank  ≤ 10 Rank  ≤ 50 Rank  ≤ 100
Left Slap 0.0200 0.0199 0.0199 0.0199 0.0199 0.0199
Right Slap 0.0197 0.0197 0.0197 0.0197 0.0197 0.0197
Left + Right Slap 0.0194 0.0194 0.0194 0.0194 0.0194 0.0194
Identification Flats 0.0360 0.0360 0.0360 0.0360 0.0360 0.0360