On the Recognition Performance of BioHashing on state-of-the-art Face Recognition models

This package is part of the signal-processing and machine learning toolbox Bob. It contains the source code to reproduce the following paper:

@inproceedings{wifs2021_biohashing_sota_face,
   title={On the Recognition Performance of BioHashing on state-of-the-art Face Recognition models},
   author={Shahreza, Hatef Otroshi and Hahn, Vedrana Krivoku{\'c}a and Marcel, S{\'e}bastien},
   booktitle={2021 IEEE International Workshop on Information Forensics and Security (WIFS)},
   pages={1--6},
   year={2021},
   organization={IEEE}
}

Installation

The installation instructions are based on conda and works on Linux systems only. Therefore, please install conda before continuing.

For installation, please download the source code of this paper and unpack it. Then, you can create a conda environment with the following command:

$ cd bob.paper.wifs2021_biohashing_sota_face

# create the environment
$ conda create --name bob.paper.wifs2021_biohashing_sota_face --file package-list.txt

$ conda activate bob.paper.wifs2021_biohashing_sota_face  # activate the environment
$ buildout

$ export wifs2021_biohashing_sota_face_path=$(pwd)

Downloading the datasets

Two datasets are used in this study which are publicly available. To download the datasets please refer to their websites:

• MOBIO
• LFW

Configuring the directories of the datasets

Now that you have downloaded the four databases. You need to set the paths to those in the configuration files. Bob supports a configuration file (~/.bobrc) in your home directory to specify where the databases are located. Please specify the paths for the database like below:

# Setup MOBIO directories
$ bob config set  bob.db.mobio.directory [YOUR_MOBIO_IMAGE_DIRECTORY]
$ bob config set  bob.db.mobio.annotation_directory [YOUR_MOBIO_ANNOTATION_DIRECTORY]

# Setup LFW directories
$ bob config set  bob.db.lfw.directory [YOUR_LFW_IMAGE_DIRECTORY]
$ bob config set  bob.bio.face.lfw.annotation_directory [YOUR_LFW_ANNOTATION_DIRECTORY]

Running the Experiments

All the pipelines used in this paper are available in the ./experiments <experiments> directory. You can run each experiment you want from the mentioned directory. For example, to run the Baseline for ArcFace on the LFW dataste, you need to run the following command in the corresponding directory:

$ cd experiments/LFW/ArcFace/Baseline/
$
$ # To run on SGE using Dask
$ $wifs2021_biohashing_sota_face_path/bin/bob bio pipelines vanilla-biometrics my_pipeline.py -l sge -o results -vvv
$ # To run locally
$ $wifs2021_biohashing_sota_face_path/bin/bob bio pipelines vanilla-biometrics my_pipeline.py -l $wifs2021_biohashing_sota_face_path/bob/paper/wifs2021_biohashing_sota_face/local.py -o results -vvv

To run all the experiments, you can use run_experiments.sh as following:

$ ./run_experiments.sh

Note: In ./run_experiments.sh file, there is a variable, run_on_grid, to execute the task on computaion grid (using dask). If you have SGE computaion grid, you can set it True, otherwise you need to set it False and run the experiments locally.

Generating ROC Plots (Figure 2)

To generate ROC plts (as in Figure 2 of the paper), you need to go to ./evaluation/ROC directory and run the bash file:

$ cd evaluation/ROC
$ ./bash.sh

Generating TMR-L_BioHash Plots (Figure 3)

To generate TMR vs L_BioHash plots (as in Figure 3 of the paper), you can use the notebook which is available in ./evaluation/TMR_LBhsh directory.

Generating Table 1

To generate Table 1 of the paper, you can use the same notebook as for Figure 3, which is available in ./evaluation/TMR_LBhsh directory.

Contact

For questions or reporting issues to this software package, contact our development mailing list.