Packages for image recognition - Robocup TU/e Robotics

Package	Build status Xenial Kinetic x64	Description
image_recognition		Meta package for all image_recognition packages.
image_recognition_util		Utils shared among image recognition packages
image_recognition_analysis		analysis for image recognition
image_recognition_msgs		Interface definition for image recognition
image_recognition_rqt		RQT tools with helpers testing this interface and training/labeling data.
tensorflow_ros		Object recognition with use of Tensorflow. The user can retrain the top layers of a neural network to perform classification with its own dataset as described in this tutorial.
tensorflow_ros_rqt		RQT tools for retraining a Tensorflow neural network.
openface_ros		ROS wrapper for Openface (https://github.com/cmusatyalab/openface) to detect and recognize faces in images.
skybiometry_ros		ROS wrapper for Skybiometry (https://skybiometry.com/) for getting face properties of a detected face, e.g. age estimation, gender estimation etc.
openpose_ros		ROS wrapper for Openpose (https://github.com/CMU-Perceptual-Computing-Lab/) for getting poses of 2D images.

Step 1: label images with the image_recognition_rqt#annotation-plugin

<-- Youtube video

Step 2: train a neural network with the tensorflow_ros_rqt

<-- Youtube video

Step 3: predict labels for new data with the image_recognition_rqt#test-plugin

<-- Youtube video

See the tutorial at openface_ros

<-- Youtube video

Clone the repo in your catkin_ws:

    cd ~/catkin_ws/src
    git clone https://github.com/tue-robotics/image_recognition.git

Build your catkin workspace cd ~/catkin_ws catkin_make