MVP v2.0.0

Details

-- Trimmed YOLO faced module for face detection
-- Insightface model is used for Face recognition
-- Totally metric learning based architecture ( will be improved upon using saimese networks)
-- Face alignment done using FAN

##How to run

pip install -r requirements.txt
$ cd faced
$ python setup.py install
$ cd ../pretrained
$ cp <download destination of insightface weights> .
$ cd ..
$ python encode/encode.py
$ python register/register.py 
$ python livestream/newmain.py

MVP v1.0.0

--Face Detection YOLO
--Face Alignment FAN
--Face Recognition VggFace
##How to run

$ cd accessfaceid-beta
$ pip install -r requirements.txt
$ cd faced
$ python setup.py install
$ cd ../training
$ mkdir weights
$ cd weights
$ cp <download destination of vggface weights> .
$ cd ../../
$ python encode.py
$ python main.py

MVP v0.0.0

-- main.py works using face_recognition and opencv
--primitive version lags in fps and compute heavy on CPU for real-time recognition and detection.
--face detection is done solely using face_recognition API
--encode.py works using face_recognition and opencv
--generates embedding of all the images in the system and then makes their encodings.