Self Morphing

This repository contains python implementation of Self_morphing with OpenCV using FERET and FRGC datasets in the GRL (Generalizable Representation Learning) paper.

Table of Contents

• GRL
  • Installation
  • Self-Morphing on FERET dataset
  • Self-Morphing on FRGC dataset
  • Citing GRL
  • Reference

Installation

1. First clone the repository

   git clone https://github.com/kashiani/SelfMorphing_landmarks.git
   

2. Create the virtual environment via conda

   conda create -n SelfMorphing python=3.7
   

3. Activate the virtual environment.

   conda activate SelfMorphing
   

4. Install the dependencies.

   pip install -r requirements.txt
   

Self-Morphing on FERET dataset

run the following command:

python Create_self_morphing_FERET.py  --source_images './src_images/FERET_MTCNN'  --output './out_images/self_FERET' --proportion_morphing 0.5 

Self-Morphing on FRGC dataset

run the following command:

python Create_self_morphing_FRGC.py  --source_images './src_images/FRGC_MTCNN'  --output './out_images/self_FRGC' --proportion_morphing 0.5 

Citing GRL

If you use this repository or would like to refer the paper, please use the following BibTeX entry

@inproceedings{kashiani2023towards,
  title={Towards Generalizable Morph Attack Detection with Consistency Regularization},
  booktitle={2023 IEEE International Joint Conference on Biometrics (IJCB)},
  pages={1--10},
  year={2023},
  organization={IEEE}
}

Reference

[1] Kashiani, Hossein, et al. "Towards Generalizable Morph Attack Detection with Consistency Regularization." 2023 IEEE International Joint Conference on Biometrics (IJCB). IEEE, 2023.