Install and Run

1. install system-level prerequisites:

sudo apt update
sudo apt install -y python-virtualenv
sudo apt install -y curl wget
sudo apt install -y libhdf5-dev libhdf5-serial-dev

2. clone repo:

git clone https://github.com/moshes7/face_fever_detection.git
cd face_fever_detection

3. create virtual env and install pkg:

virtualenv venv_27 --python=python2
source venv_27/bin/activate
pip install -r requirements.txt

4. Download (from drive) and install MXNet (.whl) with GPU support:

bash bin/get_gdrive_file.sh 1i-wgDa8rVv-9l-iR8iEhWNSLt7A9bRwZ /tmp/mxnet-1.6.0-py2-none-any.whl
pip install /tmp/mxnet-1.6.0-py2-none-any.whl

5. Download from drive and replace the file libmxnet.so in venv_27/lib/python2.7/site-packages/mxnet

bash bin/get_gdrive_file.sh 1fPVmW7MVMW1C13mv0ZEDXv1JmHRyZzbh venv_27/lib/python2.7/site-packages/mxnet/libmxnet.so

6. Download Pretrained Model: RetinaFace-R50 from drive and place in examples/RetinaFace/models

mkdir -p examples/RetinaFace/models
bash bin/get_gdrive_file.sh 1p75bDFzOa0LFTdgBJcq3BZDKOMpZ0qp4 examples/RetinaFace/models/retinaface-R50.zip
unzip examples/RetinaFace/models/retinaface-R50.zip -d examples/RetinaFace/models

7. Compile additional source:

cd examples/RetinaFace
make -j
cd -

8. Download (from drive) and install opencv (including contrib and aruco)

bash bin/get_gdrive_file.sh 1uCF6s-wjluZLIwb9150KMlfZ0G25UQiA opencv.tar.gz
tar -xzf opencv.tar.gz
echo export LD_LIBRARY_PATH=$(readlink -f opencv/lib):$LD_LIBRARY_PATH >> ~/.bashrc
deactivate
source ~/.bashrc
source venv_27/bin/activate

9. Test opencv installation (the following command should return with no errors)

python -c "import cv2.aruco"

10. Copy camera configuration files:

sudo mkdir -p /etc/eyerop
sudo cp -rf etc/* /etc/eyerop/

11. open new terminal window and run:

sudo bin/erop-proxy-cam -nproxycam-ir --perror --logmask=130

12. Run from examples/python:

 sudo LD_LIBRARY_PATH=$LD_LIBRARY_PATH ../../venv_27/bin/python  capture_opencv.py /ipc0