Evaluation Framework Information

The experimental evaluations have been conducted using the open source vein recognition framework (PLUS OpenVein SDK) provided by the University of Salzburg. This is a feature extraction and matching/evaluation framework for finger- and hand-vein recognition implemented in MATLAB. It was tested on MATLAB 2016 and should work with all version of MATLAB newer or equal to 2016. This software is under the Simplified BSD license.

The framework contains all the feature extraction and matching as well as evaluation methods used for the experiments in the paper:

Christof Kauba, Bernhard Prommegger, and Andreas Uhl "Focussing the Beam - A New Laser Illumination Based Data Set Providing Insights to Finger-Vein Recognition" InProceedings of the IEEE 9th International Conference on Biometrics: Theory, Applications, and Systems (BTAS2018), 9 pages, Los Angeles, California, USA, October 22-25, 2018.

A more detailed description of the framework as well as its sources can be found here:

PLUS OpenVein SDK

Data Set

PLUSVein-FV3 Laser/LED based Dorsal/Palmar Finger Vein Data Set

The finger-vein data set used during the evaluations is publicly available for research and non-commercial purposes and can be requested here:

PLUSVein-FV3 Database

Further informations regarding the data set can be found by following the above link.

Scores Files

General Structure of the Files

The scores files are provided as MATLAB .mat files. Each .mat file contains a struct, containing two vectors:

• positives: This vector contains the scores obtained from the genuine matches.
• negatives: This vector contains the scores obtained from the impostor matches.

The scores are all similarity scores, i.e. higher scores indicate higher similarity. Thus the genuine scores should be ideally higher than the impostor ones. The scores obtained for the 3 binary features using the Miura matcher are in the range of [0 - 0.5] while the scores obtained for SIFT are in the range of [0 - 1].

File Naming Conventions and Directory Structure

According to the experiments, there are three different sets of score files: 

• the general recognition performance evaluation of the laser and LED based scanner for the dorsal finger vein images
• the cross-scanner comparison (Laser - LED) scores
• the subset specific analysis (gender and age groups)

So the directory structure is as follows:

• Scores: contains all the scores files
• Laser: contains the scores files for the laser based scanner subset
• LED: contains the scores files for the LED based scanner subset
• Cross-Laser-LED: contains the cross-scanner comparison scores
• Subset Analysis: contains the gender- and age-group specific analysis scores
• Laser: contains the laser scanner subset analysis scores
• Age Groups: contains the age-groups scores
• Scores_GF_Age-Group-10-30: filtered scores, only containing subjects between >10 and <30 years
• Scores_GF_Age-Group-30-40: filtered scores, only containing subjects between >30 and <40 years
• Scores_GF_Age-Group-40-80: filtered scores, only containing subjects between >40 and <80 years
• Gender: contains the gender-group scores
• Scores_FE_Female: filtered scores, only containing subjects that are female
• Scores_FE_Male: filtered scores, only containing subjects that are male
• LED: contains the LED scanner subset analysis scores
• Age Groups: contains the age-groups scores
• Scores_GF_Age-Group-10-30: filtered scores, only containing subjects between >10 and <30 years
• Scores_GF_Age-Group-30-40: filtered scores, only containing subjects between >30 and <40 years
• Scores_GF_Age-Group-40-80: filtered scores, only containing subjects between >40 and <80 years
• Gender: contains the gender-group scores
• Scores_FE_Female: filtered scores, only containing subjects that are female
• Scores_FE_Male: filtered scores, only containing subjects that are male
• Settings: contains the settings files to be used with the OpenVein SDK evaluation framework to arrive at these scores based on the PLUSVein-FV3 database
• Laser: contains the settings files for the laser scanner general recognition performance evaluation
• LED: contains the settings files for the LED scanner general recognition performance evaluation
• Cross-Laser-LED: contrains the settings files for the Laser-LED scanner cross recognition performance evaluation

There are no separate settings files for the subgroup analysis (gender and age) as this analysis is based on a subset of the scores from the general recognition performance analysis.

Each of the subdirectories contains one scores file or one settings file per feature type, respectively:

• Scores_GF.mat: Scores obtained for the evaluation of the Gabor Filter based features.
• Scores_MC.mat: Scores obtained for the evaluation of the Maximum Curvature based features.
• Scores_PC.mat: Scores obtained for the evaluation of the Principal Curvature based features.
• Scores_SIFT.mat: Scores obtained for the evaluation of the SIFT based features.

For the Cross-Laser-LED settings the naming is different. These settings files are named Cross_GF.ini, Cross_MC.ini, Cross_PC.ini and Cross_SIFT.ini. Note that for the cross scanner comparison the LED scanner images are used as a gallery whereas the Laser scanner images are the probe directory.
These score and settings files can be downloaded below: