Face Recognition

Before using the scripts

You should have python 2 installed in your system and some python libraries: opencv-python 2.x and numpy. If you don't have this environment you can use docker as explained on Build and running inside docker to run the scripts. Or a dev image as explained here

Using the scripts

The face recognition only works with after face recognizer is generated.

To generate a face reconizer first you have to create a data set:

$ ./create-dataset.py $USER_ID data-set

To train the face recognizer

$ ./train.py data-set trainner

After that you can run the face reconition script

$ ./recognize-face.py

It needs that the reconizer file is located on trainner/trainner.yml

Build and running inside docker

If you don't want to set an environmet to run the scripts, you can use the provided Dockerfile to build and image and run the container:

$ docker build . -t <image-name>

To run it you have to adjust the permission to the X server host. So run the following command in the docker host machine:

$ xhost +local:root

This is a lazy workaround to permit connection from the container, and it can be improved as explained here.

After that you can run your docker container:

$ docker run -it \
    --env="DISPLAY" \
    --env="QT_X11_NO_MITSHM=1" \
    --volume="/tmp/.X11-unix:/tmp/.X11-unix:rw" \
    --device /dev/video0 \
    <image-name>

or just use the provided script:

$ ./run-dev-container.sh <image-name>

when you finish, disable permission to connect to your X server host:

$ xhost -local:root

References

All scripts are adaptions from the content explained in https://thecodacus.com/opencv-face-recognition-python-part1/