USDAN
The implementation of Unified unsupervised and semi-supervised domain adaptation network for cross-scenario face anti-spoofing, which is accepted by Pattern Recognition 2021.
An overview of the proposed USDAN method:
Congifuration Environment
- python 3.6
- pytorch 0.4
- torchvision 0.2
- cuda 8.0
Pre-training
Dataset.
Download the CASIA-FASD, Idiap Replay-Attack, and MSU-MFSD datasets.
Data Pre-processing.
MTCNN algorithm is utilized for face detection and face alignment. All the detected faces are normalized to 224$\times$224$\times$3, where only RGB channels are utilized for training. The exact codes that we used can be found here.
Put the processed frames in the path $root/processed_data
To be specific, we first utilize the MTCNN algorithm to process every frame of each video. And then, we utilize the get_files
function in the utils/utils.py
to sample frames during training. Finally, the information of selected frames are saved to the choose_*.json
file.
Data Label Generation.
Move to the $root/USDAN_*/msu_casia/data_label/
and generate the data label list:
python generate_label.py
Training
Move to the folder $root/USDAN_*/msu_casia/
and just run like this:
python train_USDAN_*.py
The file config.py
contains the hype-parameters used during training.
Testing
Run like this:
python da_test.py
Citation
Please cite our paper if the code is helpful to your researches.
@InProceedings{Jia_2021_PR_USDAN,
author = {Yunpei Jia and Jie Zhang and Shiguang Shan and Xilin Chen},
title = {Unified Unsupervised and Semi-supervised Domain Adaptation Network for Cross-scenario Face Anti-spoofing},
booktitle = {Pattern Recognition},
year = {2021}
}