Deep Anomaly Detection for Generalized Face Anti-Spoofing

This is an unofficial implemented code for paper "Deep Anomaly Detection for Generalized Face Anti-Spoofing" in pytorch.

This code works fine on our own dataset and is worth sharing.

The original paper can be find here in arxiv: https://arxiv.org/abs/1904.08241

Install

pip install torch torchvision tqdm albumentations

Usage

First, make a dir containing positive and negative folder and place the corresponding image in the folder.

Second, configure data path in dataset.py .

Then run training

python train.py

For visualization,

First, generate the txt file for t-sne visuation.

python generate_txt_for_tsne.py

Then, visuallize them

python t_sne.py

The visual effect is as shown in the figure:

Contributing

PRs accepted.

License

MIT © Aoruxue