Multi-Biometric Template Protection based on Bloom filters

Implementation of the feature level fusion of Bloom filter based protected templates proposed in [IF18] .

License

This work is licensed under license agreement provided by Hochschule Darmstadt (h_da-License).

Instructions

Dependencies

• seaborn
• numpy
• pylab
• matplotlib
• argparse

Usage

1. Run BF_template_fusion.py to fuse the Bloom filter based templates of two characteristics or instances.

   usage: BF_template_fusion.py [-h]
                                [--DB_BFtemplates_fused [DB_BFTEMPLATES_FUSED]]
                                DB_BFtemplatesA DB_BFtemplatesB fusionList
   
   Extract unprotected LGBPHS and protected Bloom filter templates from the FERET
   DB.
   
   positional arguments:
     DB_BFtemplatesA       directory where the protected BF templates for
                           characteristic A are stored
     DB_BFtemplatesB       directory where the protected BF templates for
                           characteristic B are stored
     fusionList            file where the pairs of templates to be fused are
                           specified
   
   optional arguments:
     -h, --help            show this help message and exit
     --DB_BFtemplates_fused [DB_BFTEMPLATES_FUSED]
                           directory where the protected fused BF templates will
                           be stored

   1. Input: at least 2 directories where the protected templates to be fused are stored, and a file with the list of templates to be fused. Each line of the file contains the corresponding file names separated by a blank space. It should be noted that the first characteristic should be the one comprising the biggest templates (iris in the example).
   2. Output: fused templates, stored in the directory indicated by DB_BFtemplates_fused.
   3. Two templates have been provided to show how the script works. The call should be: BF_template_fusion.py irisDB/ faceDB/ fusionList.txt

2. Run BF_template_fusion.py to fuse the Bloom filter based templates of two characteristics or instances.

   usage: computeScores.py [-h] [--scoresDir [SCORESDIR]]
                           [--matedScoresFile [MATEDSCORESFILE]]
                           [--nonMatedScoresFile [NONMATEDSCORESFILE]]
                           DB_BFtemplates matedComparisonsFile
                           nonMatedComparisonsFile
   
   Compute protected Bloom filter scores from a given DB and protocol.
   
   positional arguments:
     DB_BFtemplates        directory where the protected BF templates are stored
     matedComparisonsFile  file comprising the mated comparisons to be carried
                           out
     nonMatedComparisonsFile
                           file comprising the non-mated comparisons to be
                           carried out
   
   optional arguments:
     -h, --help            show this help message and exit
     --scoresDir [SCORESDIR]
                           directory where unprotected and protected scores will
                           be stored
     --matedScoresFile [MATEDSCORESFILE]
                           file comprising the mated scores computed
     --nonMatedScoresFile [NONMATEDSCORESFILE]
                           file comprising the non-mated scores computed

   1. Input: folders with the unprotected and protected templates (arguments can be modified at the top of the script to use other folders), as well as the folder where the scores will be stored
   2. Output: mated and non-mated scores, stored in text files with one score per row

References

More details in:

• [IF18] M. Gomez-Barrero, C. Rathgeb, G. Li, R. Raghavendra, J. Galbally and C. Busch, "Multi-Biometric Template Protection Based on Bloom Filters", in Information Fusion, vol. 42, pp. 37-50, 2018.

Please remember to reference article [IF18] on any work made public, whatever the form, based directly or indirectly on these scripts.