zhengyi144/awesome_face_antispoofing

Single Shot Face Anti-spoofing.

awesome_face_antispoofing

• This is a single shot face anti-spoofing project.
• The deep learning framework is Pytorch. Python3.5 is used.

Face landmarks

• face_alignment is used for landmarks extraction. Page face_alignment. Thanks to them.

Landmarks extraction scripts

• cd detlandmark&&python3 detlandmark_imgs.py NUAA_raw_dir

Data

• I have upload data and detected landmarks into GOOGLE DRIVE-raw.tar.gz
• You can change corresponding directory and filename in config.py
• For example train_filelists=[ ['raw/ClientRaw','raw/client_train_raw.txt',ATTACK], ['raw/ImposterRaw','imposter_train_raw.txt',GENUINE] ] test_filelists=[ ['raw/ClientRaw','raw/client_test_raw.txt',ATTACK], ['raw/ImposterRaw','raw/imposter_test_raw.txt',GENUINE] ]

Method

• Our method is straightforward. Small patched containing a face is cropped with corresponding landmarks. A binary classification network is used to distinguish the attack patches.

Training

• First, edit file config.py, choose the target network and proper batch_size.
• Then, in terminal command: make clean&&make&&python3 main.py train

Experiments

• Experiments results on NUAA[1] Image input size is as same as the imagenet.
• State-of-the-art networks are used, e.g. VGG[2], ResNet[3], DenseNet[4], Inception[5], Xception[6], DetNet[7]

Network	Acc	AUC	EER	TPR(1.0%)	TPR(.5%)
VGG-11	0.9601	0.99319872	0.038188	0.832243	0.778703
VGG-13	0.9549	0.99473387	0.036166	0.876264	0.824372
VGG-16	0.9642	0.99509249	0.035693	0.886377	0.805989
VGG-19	0.9546	0.99366445	0.041047	0.868531	0.829566
Res-18	0.9329	0.99650029	0.026037	0.927127	0.857467
Res-34	0.9568	0.99489955	0.033501	0.877933	0.791791
Res-50	0.9941	0.99981480	0.005998	0.997271	0.992861
DenseNet-121	0.9467	0.99159944	0.046996	0.806068	0.709994
DenseNet-169	0.9817	0.99859679	0.015622	0.961695	0.919752
DenseNet-201	0.9235	0.97937384	0.096862	0.798037	0.713266
Densenet-161	0.9540	0.99784970	0.020363	0.953331	0.894706
Inception-V3	0.9114	0.98636584	0.061354	0.693337	0.583299
Xception	0.9841	0.99929783	0.012338	0.983641	0.970793
DetNet	0.9685	0.99998105	0.001562	1.000000	0.999703

Reference

• [1]Tan X, Li Y, Liu J, et al. Face liveness detection from a single image with sparse low rank bilinear discriminative model[C]// European Conference on Computer Vision. Springer-Verlag, 2010:504-517.
• [2]Simonyan K, Zisserman A. Very deep convolutional networks for large-scale image recognition[J]. arXiv preprint arXiv:1409.1556, 2014.
• [3]He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2016: 770-778.
• [4]Huang G, Liu Z, Van Der Maaten L, et al. Densely connected convolutional networks[C]//CVPR. 2017, 1(2): 3.
• [5]Szegedy C, Vanhoucke V, Ioffe S, et al. Rethinking the inception architecture for computer vision[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2016: 2818-2826.
• [6]Chollet F. Xception: Deep learning with depthwise separable convolutions[J]. arXiv preprint, 2017: 1610.02357.
• [7]Li Z, Peng C, Yu G, et al. DetNet: A Backbone network for Object Detection[J]. arXiv preprint arXiv:1804.06215, 2018.