You need python and cmake:
sudo apt install python cmake v4l2loopback-dkms libv4l-dev
Load project deps:
pip install -r requirements.txt
Download and uncompress pretrained face model:
wget http://dlib.net/files/shape_predictor_68_face_landmarks.dat.bz2
bzip2 -d shape_predictor_68_face_landmarks.dat.bz2
python3 preview.py open.jpg video.webm out.avi --vertical
Enable and create eye opener camera:
sudo modprobe v4l2loopback card_label="Eye contact" exclusive_caps=1 max_buffers=2
This shoud create new video device (/dev/video1 usually). Check it:
v4l2-ctl --list-devices
Capture your open eyes photo from real camera:
python3 capture.py open.jpg /dev/video0
Runt it:
python3 eye_contact.py open.jpg /dev/video0 /dev/video1