De-styled secure face recognition system

1. Introduction

This is a development example of secure face recognition system that front and rear end separation . It is based on the face recognition package of insightface and the face anti-spoofing solution of SSAN

2. Requirement

• Python 3.6+
• npm 8.1.0+
• cuda 11.0+

3. Usage

1. Install the dependencies

pip install -r requirements.txt
cd ./front/face_front
npm install

2.Start the server

cd ./front/face_front
npm run dev
cd ../../
cd ./back
python main.py

3. Open the browser and enter the address

http://localhost:5173

4. Demo

1. Face slice

Use the face_slice to get the face image from the source image.We get the landmarks of faces before we get this.

2. Face recognition

3. Face compare

4. Face management

it can add,edit,delete the pre-existing face

5. Use webCamera

5. Future

• warning when the spoof appeared
• enable to use the local camera
• better performance