fingermatchR

fingermatchR provides fingerprint matching tools based on NIST’s MINDTCT and BOZORTH3 algorithms, as released in NIST’s Biometrics Image Software suite (NBIS). It also provides R bindings to the FingerJet minutiae extraction tool and to libfmr to parse ANSI/INCITS 378-2004 and ISO/IEC 19794-2 fingerprint minutiae records.

Installation

Make sure you have NBIS installed with binaries available in PATH. Check in particular that you have the mindtct and bozorth3 executables installed and available:

$ mindtct -version && bozorth3 -version
Standard Version: ANSI/NIST-ITL 1-2007
NBIS Non-Export Control Software Version: Release 5.0.0
...

You can install the development version of fingermatchR from GitHub with:

if (!require(devtools)) install.packages("devtools")
devtools::install_github("forensic-science/fingermatchR")

Examples

library(fingermatchR)
library(dplyr)

Optionally, define location of NBIS executables:

# This is required if NBIS executables are not available in PATH.
#options(NBIS_bin = "/home/olivier/Desktop/Research/forensic-science/NBIS/src/bin")
options(NBIS_bin = "/usr/local/NBIS/bin")

Minutiae detection

Detect minutiae using mindtct:

imgfiles = list.files("data-raw", pattern = "*.png", full.names = TRUE)

# Raw mindtct output
out = mindtct(imgfiles, outputdir="data-raw")
#> ℹ Running mindtct on 3 image files...
#> ✓ done running mindtct.

# Tidy minutiae descriptions
minutiae = tidyMinutiae(out)

Plot fingerprint image and its binarization with detected minutiae:

# Plain and rolled index fingerprints for the same individual
plotMinutiae(out[1:3, ])

Match scores

Compute pairwise fingerprint match scores:

matchscores(out, outputdir = "data-raw")
#> # A tibble: 3 x 3
#>   score probe_index gallery_index
#>   <chr>       <int>         <int>
#> 1 3               1             2
#> 2 31              1             3
#> 3 7               2             3

Compute all-to-all fingerprint match scores between two lists:

matchscores(out, out, outputdir = "data-raw")
#> # A tibble: 9 x 3
#>   score probe_index gallery_index
#>   <chr>       <int>         <int>
#> 1 194             1             1
#> 2 3               1             2
#> 3 31              1             3
#> 4 3               2             1
#> 5 512             2             2
#> 6 7               2             3
#> 7 31              3             1
#> 8 7               3             2
#> 9 622             3             3

Using the FingerJet minutiae extration tool

Detect minutiae using FingerJet:

imgfiles = list.files("data-raw", pattern = "*.png", full.names = TRUE)

# Raw mindtct output
fingerjet = fj_minutiae(imgfiles, outputdir="data-raw/fingerjet")
#> Running FingerJet minutiae extrator on 3 image files.

Plot fingerprint image and its binarization with detected minutiae:

# Plain and rolled index fingerprints for the same individual
plotMinutiae(fingerjet[1:3, ])

Compute pairwise fingerprint match scores:

matchscores(fingerjet, outputdir = "data-raw/fingerjet")
#> # A tibble: 3 x 3
#>   score probe_index gallery_index
#>   <chr>       <int>         <int>
#> 1 0               1             2
#> 2 22              1             3
#> 3 11              2             3

References

• Watson et al. (2007) User’s Guide to Export ControlledDistribution of NIST Biometric Image Software (NBIS-EC). National Institute of Standards and Technology.
• Watson et al. (2007) User’s Guide to NIST Biometric Image Software (NBIS). National Institute of Standards and Technology.