PatchNet: A Simple Face Anti-Spoofing Framework via Fine-Grained Patch Recognition

This repository implements PatchNet from paper PatchNet: A Simple Face Anti-Spoofing Framework via Fine-Grained Patch Recognition

Reference

[1] PatchNet: PatchNet: A Simple Face Anti-Spoofing Framework via Fine-Grained Patch Recognition
[2] CDCN repository: CDCN-Face-Anti-Spoofing.pytorch

Project Structure

Implementation-patchnet
|
|---config
| |--config.yaml
|
|---dataset
| |--FAS_dataset.py
| |--transform.py
|
|---engine
| |--__init__.py
| |--base_trainer.py
| |--Patchnet_trainer.py
|
|---metrics
| |--losses.py
| |--meter.py
|
|---models
| |--CDCNs.py
| |--convnext_tiny.py
| |--DC_CDN.py
| |--resnet18.py
| |--swin_base.py
|
|---tool
| |--test.py
| |--train.py
|
|---utils
| |--utils.py
|
|---README.md
|---requirements.txt

Installation

$ python3 -m venv env
$ source env/bin/activate
$ pip install -r requirements.txt

Usage

Data preparation

datasets
|---images
| |--img1
| |--img2
| |...
|---train.csv
|---val.csv
|---test.csv

with [set_name.csv] have format (label only has 2 class: 0-Spoofing, 1-Liveness): \

image_name | label
img_name1 | 0
img_name2 | 1
...

training

python3 train.py

Testing

Go to tool/test.py and fix saved_name to your path to checkpoint
Run

python3 test.py

Contributer

Tien Thong Doan
Minh Chau Nguyen
Minh Hung Nguyen