Parallel framework for fingerprint identification.

This package contains the source code for the framework described in the following papers:

• D. Peralta, I. Triguero, R. Sanchez-Reillo, F. Herrera, J.M. Benítez. Fast Fingerprint Identification for Large Databases. Pattern Recognition 47:2 (2014) 588–602. doi: 10.1016/j.patcog.2013.08.002
• D. Peralta, I. Triguero, S. García, F. Herrera, J.M. Benítez. DPD-DFF: A Dual Phase Distributed Scheme with Double Fingerprint Fusion for Fast and Accurate Identification in Large Databases. Submitted to Information Fusion.

mpi-afis loads a set of template minutiae files (xyt or ISO 19794-2 formats), and iteratively loads input minutiae files, looking them up in the template database. Two matching algorithms are implemented, as well as several configuration options for the framework. mpi-afis is flexible in the number of processes and threads it executes; both can be specified in the parameters to suit the underlying hardware.

Two pairs of fingerprints are provided as an example.

The project can be compiled by executing:

make

Requirements:

• MPI (tested with OpenMPI 1.8.7), ideally a thread-safe compilation.
• OpenMP

Example executions:

./DPDDFF -h
mpirun ./bin/genericMatching -t example_fps/test_template.dat -i example_fps/test_input.dat
mpirun -np 2 ./bin/genericMatching -a mcc -k 10 -s m -t template_files.dat -i input_files.dat -N 8 -C LSSR
mpirun -np 2 ./bin/DPDDFF -a jiang,mcc -k 10 -s r -r 10 -t template_files_F1.dat,template_files_F2.dat -i input_files_F1.dat,input_files_F2.dat -v DD