This project is aimed to recognize a person. it uses CNN(convulotional neural network) for this purpose. The input data comes from camera.
Prerequisites: ROS Kinetic --> http://wiki.ros.org/kinetic/Installation/Ubuntu Tensorflow > 1.80 --> https://www.tensorflow.org/install/ needed packages such as numpy, opencv, ...
building the package: -cloning the project from https://gitlab.com/MRL-AtHome/person_recognition -making a ros workspace --> $ mkdir -p ~//src $ cd ~// $ catkin_make $ source devel/setup.bash for more information go to http://wiki.ros.org/catkin/workspaces -cloning the project in the workpace from https://gitlab.com/MRL-AtHome/person_recognition -building the project --> $ catkin_make $ catkin_make install -source the environment setup for ros --> $ source /opt/ros/kinetic/setup.bash -source the environment setup for the package --> $ source %YOUR_WORKSPACE_LOCATION%/devel/setup.bash
running the nodes: -runnig the roscore node --> $ roscore -running the package nodes --> $ rosrun person_recognition image_publisher.py $ rosrun person_recognition start.py -run ros 3D visualizer --> & rosrun rviz rviz
now you can see the input and output image. you can add a name to the recognized face --> $ rosservice call person_recognition "name:['%THE_NAME%']"
