Engagement Detector

ROS package to detect overall users engagement from a robot's ego camera in human-robot interactions.

The approach is detailed in the journal paper: Del Duchetto F, Baxter P and Hanheide M (2020) Are You Still With Me? Continuous Engagement Assessment From a Robot's Point of View. Front. Robot. AI 7:116. doi: 10.3389/frobt.2020.00116

Install

Install package with dependencies

1. Install python catkin util: pip install catkin_pkg

2. In a terminal go into the root folder of the package: cd engagement_detector/

3. and install: pip install .

4. Then download the keras model of the newtork: ./download_model.sh

5. Now, you can build the package in your catkin workspace (i.e. http://wiki.ros.org/catkin/Tutorials/create_a_workspace).

Launch

roslaunch engagement_detector engagement_detector.launch

Engagement value

The predicted engagment value is published on the topic /engagement_detector/value.

Launch parameters

• image: (default /camera/color/image_raw) input image
• debug_image: (default: true) whether to publish the out debug image
• out_image: (default: /engagement_detector/out_image) the debug image topic

Real-time visualization of engagement

rosrun image_view image_view image:=/engagement_detector/out_image will show the camera image with the engagement serie data plotted on it. Something like this:

Single user example	Multi users example

On a GeForce GTX 1060 6GB GPU the engagement value is published at a rate of about 5 hz.