6 FpVTE 2012—Class C

Results in this section involve different impression types with all ten fingers. When plain impression fingerprints were used, the configuration contained left slap, right slap, and left and right thumbs (i.e., finger positions 11–14, or a 4-4-1-1 configuration².), as both probes and references. Probes were searched against enrollment databases of 5 000 000 subjects containing all ten fingers.

The datasets in this section are equivalent to those used in NIST FpVTE 2012 (Class C). 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 for the plain impressions, was performed by hisign+000C.

6.1 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 6.1. Each template was generated by a single function call providing hisign+000C all of the listed image types.

Table 6.1: Summary of template generation for FpVTE 2012—Class C.

Image Contents	Template Type	Failure to Extract	≈Total
Ten Fingers (Plain)	Probe	290	30 000
Ten Fingers (Roll)		0	30 000
Ten Fingers (Plain)	Reference	82 507	5 000 000
Ten Fingers (Roll)		0	5 000 000

6.2 Search

The probe templates from Table 6.1 were searched against two enrollment databases of 5 000 000 subjects containing images as specified in the reference template rows of Table 6.1. Approximately one-third of the probes had a corresponding mate in the enrollment database.

6.2.1 DET

The DET plot in Figure 6.1 shows the tradeoff of errors of hisign+000C when searching probe templates from FpVTE 2012—Class C against enrollment databases of 5 000 000 subjects created from the reference templates from FpVTE 2012—Class C where, for approximately one-third of the probes, a single mated identity was present. Tabular versions of FNIR at select FPIR can be viewed in Table 6.2.

Figure 6.1: DET when searching probe templates from FpVTE 2012—Class C against a enrollment databases generated from reference templates from FpVTE 2012—Class C.

Table 6.2: FNIR values from the DET plotted in Figure 6.1.

Probe Content	Reference Content	FPIR  ≤ 0.001	FPIR  ≤ 0.005	FPIR  ≤ 0.01
Ten Fingers (Plain)	Ten Fingers (Plain)	0.0281	0.0275	0.0275
Ten Fingers (Plain)	Ten Fingers (Roll)	0.0102	0.0102	0.0102
Ten Fingers (Roll)	Ten Fingers (Roll)	0.0006	0.0006	0.0006

Table 6.3: Similarity score thresholds from the DET plotted in Figure 6.1.

Probe Content	Reference Content	FPIR  ≤ 0.001	FPIR  ≤ 0.005	FPIR  ≤ 0.01
Ten Fingers (Plain)	Ten Fingers (Plain)	1 045.75	864.0000	799.0000
Ten Fingers (Plain)	Ten Fingers (Roll)	791.50	760.0000	749.3125
Ten Fingers (Roll)	Ten Fingers (Roll)	818.50	792.6875	780.0000

6.2.2 CMC

The CMC plot in Figure 6.2 show the FNIR of hisign+000C when searching probe templates from FpVTE 2012—Class C against enrollment databases of 5 000 000 subjects created from the reference templates from FpVTE 2012—Class C where, for approximately one-third of the probes, a single mated identity was present. Tabular versions of FNIR at select ranks can be viewed in Table 6.4.

Figure 6.2: CMC when searching probe templates from FpVTE 2012—Class C against enrollment databases generated from reference templates from FpVTE 2012—Class C.

Table 6.4: FNIR values from CMC plotted in Figure 6.2.

Probe Content	Reference Content	Rank 1	Rank  ≤ 2	Rank  ≤ 5	Rank  ≤ 10	Rank  ≤ 50	Rank  ≤ 100
Ten Fingers (Plain)	Ten Fingers (Plain)	0.0298	0.0279	0.0275	0.0275	0.0275	0.0275
Ten Fingers (Plain)	Ten Fingers (Roll)	0.0125	0.0104	0.0103	0.0102	0.0102	0.0102
Ten Fingers (Roll)	Ten Fingers (Roll)	0.0034	0.0008	0.0006	0.0006	0.0006	0.0006

---

2. The FRIF TE API remaps finger positions 11 and 12 to 1 and 6, respectively↩︎