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)

STEP

1.run API.py: take a photo (close index and thumb finger to select and move ; close index and middle finger to select and push these two finger down together to realize clic on the botton) 2.move the generated image to new folder in folder FEATURE named "./square" 3.run batch_extractor.py to get the relative feature which saved as numpy 4.to test the picture you have taken is ok or not, you can run dissimilarity.py before that, you need to make sure "features" are in your generated folder "./square_features" 5.move "features" to folder SERVER and change the path in FHE.py 6.run FHE.py

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