This is the server side implementation of the SPECTRE for ITSC 6166/8166 Computer Communivation and Networks.
This is another implementation for the original project. In this implementation, you are able to run a service in the workstation side and run a image collection in the robot/client side.
The server side would be a workstation it will be used for deploy the trained model. The code is tested in the ROS2 foxy envrionment. Please make sure you have ROS2 foxy installed.
Similar to the original project, you need to install all the required libaries/packages. The suggested way is to using a conda virtual environment. You can follow the steps to make the service runing:
step 1 Create a virtual env
Conda create -n spectre8 python=3.8
step 2 Make a workspace, clone the code and install the requirements
git clone https://github.com/YanzeZhang97/SPECTRE_server_ws.git
cd SPECTRE_server_ws
pip install -r requirement.txt
step 3 Install additional requirements
conda install pytorch3d -c pytorch3d
cd python_service/python_service/external/face_alignment
pip install -e .
cd ../face_alignment
git lfs pull
pip install -e .
step 4 Download the pretrained model
cd ../../../../
bash quick_install.sh
colcon build
source install/setup.bash
ros2 run python_service pythonservice
To see the outputs, please run
ros2 run rqt_image_view rqt_image_view