NIST logo


Biometric Special Databases and Software

This is a listing of test data produced by the Image Group for use in evaluating automated OCR, fingerprint classification/matching, and face recognition systems. Documentation files are postscript format compressed with the UNIX compress utility. The approximate sizes of the compressed files are indicated in square brackets.




Multiple Biometrics


Newest Items

  • Special Database 27A - Finger Minutiae from Latent and Matching Tenprint Images for 1000ppi imagery and has been made available for electronic distribution. This database is updated from Special database 27 which provided 500ppi imagery to research community. 
    Documentation [3.6K] Text File, sd27a.7z  [4 GB] 7-Zip Compressed archive


Fingerprint Databases



  • Special Database 10 - NIST Supplemental Fingerprint Card Data (SFCD) for NIST Special Database 9 with include rolled fingerprints and harder to find fingerprint classifications like arch, tented arch and low count loops.
    documentation [1076K]


  • Special Database 14 - NIST Mated Fingerprint Card Pairs 2.
     Has 2,700 ten-print card pairs of rolled fingerprints (no plain impressions).
    documentation [1,020K]






  • NBIS - NIST Biometric Image Software


  • PCASYS - NIST Pattern-level Classification Automation System for Fingerprints.
    Included on NBIS CD-ROM


  • NFIS2 - NIST Fingerprint Image Software Version2.
    Included on NBIS CD-ROM



Mugshot/Face Databases



Multiple Biometrics Databases

  • Biometric Scores Set (BSSR1) - a set of raw output similarity scores from two c. 2002 face recognition systems and one c. 2004 fingerprint system, operating on frontal faces, and left and right index live-scan fingerprints, respectively. The release includes true multimodal score data, i.e. similarity scores from comparisons of faces and fingerprints of the same people.


  • The One-to-many Multimodal Fusion Package - matching results from algorithms submitted to IREX III and MBE 2010. The package is intended to facilitate research into one-to-many methods of score-level fusion.


OCR Databases











Bookmark and Share