hri_face_body_matcher

The hri_face_body_matcher package

README

hri_face_body_matcher

Overview

hri_face_body_matcher is a ROS4HRI-compatible face to body matcher node.

It finds the most likely matches between the recognized faces and bodies based on their relative position in the source image.

Algorithm

For each of the possible associations of recognized face and body, a matching cost is computed, linearly decreasing with the distance between the body nose and the face center in the image.

The rate the confidence drops is proportional to the ~confidence_scaling_factor parameter and the face size, intended as its diagonal length.

confidence = max(0, 1 - \frac{distance * c.s.f.}{2 * face\ size})

ROS API

Parameters

All parameters are loaded in the lifecycle configuration transition.

  • ~confidence_threshold (double ∈ [0,1], default: 0.5): Candidate matches with confidence lower that this threshold are not published.

  • ~confidence_scaling_factor (double > 0, default: 2.0): Factor scaling how quickly the estimated confidence drops as the distance between the matched face and body increases.

Topics

This package follows the ROS4HRI conventions (REP-155). If the topic message type is not indicated, the ROS4HRI convention is implied.

Subscribed

  • /humans/bodies/tracked

  • /humans/bodies/<body_id>/skeleton2d

  • /humans/faces/tracked

  • /humans/faces/<face_id>/roi

Published

  • /humans/candidate_matches

Execution

ros2 launch hri_face_body_matcher hri_face_body_matcher.launch.py

Example

For an example of usage, execute in different terminals:

  • USB camera:

    1. apt install ros-humble-usb-cam

    2. ros2 run usb_cam usb_cam_node_exe

  • HRI face detect:

    1. Either

    • if you are on a PAL robot apt install ros-humble-hri-face-detect

    • otherwise build and install from source.

    1. ros2 launch hri_face_detect face_detect.launch.py

  • HRI fullbody:

    1. Either

    • if you are on a PAL robot apt install ros-humble-hri-fullbody

    • otherwise build and install from source.

    1. ros2 launch hri_fachri_fullbody hri_fullbody.launch.py

  • HRI face body matcher:

    1. apt install ros-humble-hri-face-body-matcher

    2. ros2 launch hri_face_body_matcher hri_face_body_matcher.launch.py

  • RViz with HRI plugin:

    1. apt install ros-humble-rviz2

    2. Either

    • if you are on a PAL robot apt install ros-humble-hri-rviz

    • otherwise build and install from source.

    1. rviz2

In RViz, add the ‘Humans’ plugin to see the detected faces. The face IDs should be permanently assigned to the same people.