QSRlib

QSRlib is a library that allows the computation of Qualitative Spatial Relations and Calculi, as well as a development framework for rapid implementation of new QSRs.

Library

QSRLib concept

Developed by
License
MIT license (MIT)

Main Characteristic

QSRlib is a library that allows the computation of Qualitative Spatial Relations and Calculi, as well as a development framework for rapid implementation of new QSRs. The aims of QSRlib are to:

  • provide a number of Qualitative Spatial Representations (QSRs) that are well known, and in common use in the scientific community;
  • expose these QSRs via a standard IO interface that allows quick and easy re-usability, including a ROS interface to allow use in cognitive robotic systems;
  • provide a flexible and easy to use infrastructure that allows rapid development of new QSRs that extend the library;
  • deliver abstracted QSRs overtime in an aggregated representation that facilitates further inference.
Research areas
Collaborative AI Explainable AI Integrative AI Physical AI
Technical Categories
Automated reasoning Commonsense reasoning Computer vision Knowledge Representation Machine learning Robotics and automation
Last updated
12.06.2021 - 18:19
Detailed Description

Language of the library: Python

More information: A typical usage of QSRlib would be an intelligent system, such as a robot for example, which acquires visual data via an RGB-D camera, such as a Kinect, and via object recognition and skeleton tracking is able to perceive the individual entities in the world. The system can then make calls to QSRlib in order to abstract this input data and form a qualitative representation of the perceived world scene. This could then be used to recognise activities in natural scenes using already learnt models expressed using QSRs in the QSRlib library.

QSRlib provides:

  • a Python library that can be imported to any Python application;
  • a number of QSRs that are well known to the scientific community;
  • tandard IO interface (including a ROS interface) for quick use of QSRs;
  • tools to allow rapid development of new QSRs that extend the library;
  • abstracted QSRs overtime in an aggregated representation that facilitates further inference.

QSRlib has been used in various research and teaching projects, and it is available as a standalone python package or as a ROS package. This resource is a collaborative work by the Universities of Leeds, Lincoln and Birmingham (see list of authors below) and has received funds from the EC under FP7 project STRANDS, 600623. QSRlib is released under MIT license. See further details in https://qsrlib.readthedocs.io/en/latest/rsts/handwritten/license.html.

External links:

Trustworthy AI
QSRlib is a library that facilitates the use and development of QSR. QSRlib has been safely used in various research and teaching projects. Notice however that safety and trustworthy AI should be guaranteed by the user and designer. As stated in the license, this software is provided "as is", without warranty of any kind, express or implied, including but not limited to the warranties of merchantability, fitness for a particular purpose and noninfringement. In no event shall the authors or copyright holders be liable for any claim, damages or other liability, whether in an action of contract, tor or otherwise, arising from out of or in connection with the software or the use or other dealings in the software.
GDPR Requirements
QSRlib library does not collect or use any personal data from users, hence GDPR does not apply.