Ros2 synthetic facial generation

[ekakalet]

This is a pull request for possible merging of the following branches (ros2, ro2_synthetic_facial_generation). The branch ro2_synthetic_facial_generation includes a node in ros2 for creating multi-view facial synthetic images.

[ekakalet]

This PR is ready for review

[tsampazk]

Thank you @ekakalet for the new package and node!

There are several issues that need to be resolved before merging this pull request. Could you please try to run the node locally (you can find all the relevant instructions for running in #256) and making the required changes to run it in addition to the changes relating to the comments i left?

[ekakalet]

Thanks for the fixes so far. Did you manage to run the node locally? The algorithm imports don't work as they have wrong paths according to the existing directories and even after fixing them, the project path is not importable.

Can you provide some specific guidelines to run the node on an example image or similar, at least for me to test it.

@tsampazk

The node successfully executed.
“export PYTHONPATH=$OPENDR_HOME:$PYTHONPATH”
Following the instruction on https://github.com/opendr-eu/opendr/blob/master/projects/data_generation/synthetic_multi_view_facial_image_generation/README.md for preparing the execution of the main module. Arranging the input parameters (pitch and yaw angles etc.) or keeping the default, the ROS2 node will be execute with the following
“ ros2 run data_generation synthetic_facial_generation ”
with the prerequisite of existence a subscription on video (image sequence) topic
“ input_rgb_image_topic="/image_raw" ” depicting a facial image.
The node will be executed for the first 10 frames of the video input topic extracting in the output topic the multi-view facial images.

[tsampazk]

Thanks, just some final minor issues.

[tsampazk]

I think we can go ahead and merge this, thanks!