Data Enclave Guidelines

Procedures for accessing data enclave at MSU-BRIC are described below:
  • To access the data enclave, users would need to visit MSU-BRIC facility on MSU campus. Remote access to the data enclave is not allowed.
  • Adequate lead time is needed for MSU-BRIC to prepare for the visit. MSU-BRIC will work with the user to facilitate their visit.
  • A linux environment (CentOS 6.8) is provided with pre-installed GCC 4.4.7 compiler with support for C++0x. The user would give MSU-BRIC staff any code they want to be moved to the enclave and members of the team could move those. The staff will check the integrity of the user’s code. The user, once present in MSU-BRIC premises, will be assigned a terminal, able to establish a SSH session to the enclave, make modifications to their code if needed, and generate the similarity scores. Images can neither be added nor removed from the data enclave. No internet access will be provided at the terminal for data security.
  • The path to the fingerprint images parent directory will be provided to the user. The structure of the directory is as follows:
  • Only pairwise similarity scores between the images stored in the Data Enclave can be output by the program in the designated directory, assigned by the MSU-BRIC staff. The similarity matrix, representing a set of pairwise comparisons between two lists of fingerprints, should be written in the format used in NIST MBGC, outlined on pages 12-13 of the following document: https://github.com/biometrics/openbr/blob/master/docs/docs/misc/MBGC_file_overview.pdf
  • The output produced by the user’s job will be examined by MSU-BRIC staff to ensure that neither the fingerprint images nor fingerprint features extracted from the images are being retrieved.
  • Per the terms of data use agreement, following paper must be cited in any publication that results from the use of this data: S. Yoon and A.K. Jain, "Longitudinal Study of Fingerprint Recognition", Proc. National Academy of Sciences (PNAS), Vol. 112, No. 28, pp. 8555-8560, July 2015.
  • MSU reserves the right to add/modify the Data Enclave Guidelines, if needed, to ensure the safety of data.


    • The following procedure, as listed in the Yoon and Jain paper, should be followed for generating the similarity matrix:

  • Each of the 15,597 subjects has at least five 10-print cards, providing 122,685 10-print cards in total.
  • Genuine Scores Generation: For each subject i with ni fingerprint impressions, all pairwise comparisons are conducted; i.e., (ni choose 2) genuine match scores are generated. For each finger position, 481,181 genuine match scores are computed.
  • Impostor Scores Generation: The impostor scores are obtained by comparing subject i’s ni fingerprint impressions to the first impressions from all other subjects. For each finger position, 1,913,395,260 impostor match scores are obtained.
  • The total number of the generated similarity scores for all 10-fingers is approx. 20 billion, requiring 72 GB storage. The user is responsible for bringing an external storage to store the similarity matrix output.
To request access to the fingerprint images stored in the MSU-BRIC enclave, please fill and submit the form below:

Request to access MSU/MSP Longitudinal Fingerprint Data Enclave