SuperControlSystem

Face_Recognization (FaceNet)

面部识别 (FaceNet)

Fully Homomorphic Encryption over the Torus (HNP)

环面全同态加密 (TFHE)

Hand_Face_Detection (Mediapipe)

手脸检测 (Mediapipe)

DEMO (pesudo)

BOTTON

• focus on your face (Blue one on the right side)
• take a photo but you need to activate the blue botton at first (grey one the the right side)
• login (green one on the left side)
• logout (red one on the left side)

IT CAN

• FACE FEATURE EXTRACTOR
• FULLY HOMOMORPHIC ENCRYPTION
• SYSTEM WITHOUT TOUCH
• LIVE DETECTION (BLINK DETECTION)

VERSION OF PYTHON

FEATURE API python3.7 SERVER python3.8

TIPS

• weight document already include
• path may need to change (relative path is in used, normally not necessary to change)
• image needs be prepared by yourself (upload or use API.py to take pictures)
• image should be square and use same clear light
• API.py can cut a almost square:

189|   square = self.image[210:210+400,450:450+380] # change this line to get perfect square

• requirement.txt is ready and I suggest you to install in your virtual environnement:

pip install requirements.txt # to get all package (some package is not necessaire)

• number of photo for each clic on botton can be changed:

404|   photo = 6 # to take 5 picture each clic 

HOW TO RUN

python API.py [--user]
python batch_extractor.py # in the folder FEATURE, need to add your dataset folder 
python FHE.py # in the folder SERVER, need to add your feature's numpy doc

RESULT

FHE

FACENET