lizhi-ntu / ockd-facepad

This repo provides PyTorch implementation of the paper One-Class Knowledge Distillation for Face Presentation Attack Detection to appear on IEEE Transactions on Information Forensics & Security (TIFS).

🌼OCKD-FacePAD

This repo provides PyTorch implementation of the paper One-Class Knowledge Distillation for Face Presentation Attack Detection to appear on IEEE Transactions on Information Forensics & Security (TIFS).

🍀 Data Preparation

1. Please request and download the datasets. You may use the following address:

   🌏 NTU ROSE-YOUTU

   🌏 CASIA FASD

   🌍 IDIAP REPLAY-ATTACK

   🌎 MSU MFSD

   🌍 OULU-NPU

2. Please install dlib(19.24.0) and opencv(4.5.5) in your anaconda environment.

3. Please download the pretrained shape predictor model from here.

4. Please use preprocessing.py to get face images from videos. The prepocessed data for client-specific one-class domain adaptation setting are available here.

5. Please find the data division of the challenging experimental setting here.

🍀 Teacher Network Training

👀 Please use the example script train_teacher.py to train the Teacher Network.

🍀 Student Network Training

👀 Please use the example script train_student.py to train the Student Network.

🍀 Model Evaluation

👀 You may use the example script evaluation_student.py to evalute the pretrained model.

🍀 Others

👀 Everything in this repo can NOT be used for commercial purpose.

👀 If you have any questions, feel free to open an issue or contact me via email.

👀 The implementation of sparse learning in our codes is based on library.

👀 If you use this repo in your work, please use the following citation.

@ARTICLE{9782427, author={Li, Zhi and Cai, Rizhao and Li, Haoliang and Lam, Kwok-Yan and Hu, Yongjian and Kot, Alex C.}, journal={IEEE Transactions on Information Forensics and Security}, title={One-Class Knowledge Distillation for Face Presentation Attack Detection}, year={2022}, volume={}, number={}, pages={1-1}, doi={10.1109/TIFS.2022.3178240} }

ockd-facepad's Issues

怎么生成二分类得分文件

我想将测试集的得分生成到txt文件，可是我看代码中是求得相似度

how to run the entire project ?

I am facing an issue to run the project. Can you please tell me the steps to run the project

bad result

Hi,I am interested in your method but I found that the result can't catch up with paper. Is the checkpoint final version?

About Similarity Score

Hi, Can I ask a question? The higher similarity means the input have a high probability of real sense?