Robot App Store

Robot App™ Description

Developer: Arturo Bajuelos

Detects and tracks a face in the image and publishes it.

The qbo_face_tracking package is part of the qbo_face_vision ROS stack, and contains an implementation of node that detects and tracks a face in the input image via computer vision algorithms.

It subscribes to an image topic and publishes both the face image and a message with its position and approximate distance from the camera. You can also change its several ROS Parameters to configure the algorithm behavior.

Installation Instructions

Delivered as:

ROS Package

You can download the [qbo_face_vision] Then, simply build the stack and its dependencies:

git clone https://github.com/OpenQbo/qbo_face_vision
rosmake --rosdep-install qbo_face_vision

Robot App Requirements

Software

ROS

Hardware:

Q.bo robot

Other requirements:

Internet connection

Usage Instructions

Run qbo_face_tracking node

You can either run qbo_face_tracking node directly:
rosrun qbo_face_tracking qbo_face_tracking

Or use the launcher “qbo_face_tracking.launch” located in “launch/” folder:
roslaunch qbo_face_tracking qbo_face_tracking.launch

Observe detected faces
You can see the detected faces by running the image_view node in another console and subscribing to the qbo_face_tracking/viewer topic.

Subscribe to the qbo_face_tracking/viewer uncompressed images:
rosrun image_view image_view image:=/qbo_face_tracking/viewer

Subscribe to the qbo_face_tracking/viewer compressed images (must have image_pipeline stack installed):
rosrun image_view image_view image:=/qbo_face_tracking/viewer theora