Virtual Worlds for Machine Learning


This course will take place as an online event. The link to the streaming platform will be provided to the registrants only.

The Unreal Engine is one of the state-of-the-art 3D rendering engines, mainly used for game development. In recent years, however, its use in industry and science has been steadily increasing, which is further supported by new features from the producer Epic Games Inc.

This course explores the visualization of models, particularly Functional-Structural Plant Models, to improve experimental pipelines as well as the scalability of Deep Neural Networks. Synthetic data has already proven to be successful in overcoming certain aspects of data scarcity. This course will primarily shed light onto the use of the Synavis Framework for plant sciences, the use of synthetic data as well as the combination of modeling virtual worlds with experimental data acquisition. Participants will learn what synthetic data is, what it might be used for, and how it affects the training of neural networks. The course will also give an overview of research projects using synthetic data as well as technical aspects of generation. Moreover, the course will touch on digital twin modelling and best-practices on how realistic synthetic data must be to be of use.


  • Introduction to Computer Vision
  • Introduction to ML for Plant Science
  • Artifacts in Plant Scientific IMaging
  • Previous work and introduction to Synavis
  • Plant Modelling and Digital Twins
  • INVITED TALK Arnaud Bouvry
  • ---------------------------------------------
  • INVITED TALK Guillaume Lobet
  • INVITED TALK Daniel Weißen
  • Synthetic Data Pipelines for Scientific Use-Cases
  • How Realistic must synthetic data be?
  • Modelling through Synavis
  • LiDAR measurements from Unreal Engine
  • What measures mean in different contexts
  • How to increase data variability


  • Basic knowledge of Machine Learning
  • This course should be visited with previous knowledge of Unreal Engine. While it provides some insight into how Unreal Engine work, it is not to be understood as an introduction. We refer to our in-house course “Introduction to Unreal Engine”, which typically takes place in March each year.



This course is given in English.


21-22 May 2024, 09:00-12:00, 13:00-16:00

!! Course dates changes !!


Online: via Zoom.

Further information:

please visit the the document


Dirk Helmrich, JSC

    • 1
      Visualization Pipelines with Unreal Engine

      Scalability, Generalization, Domain Visualization

      Speaker: Dirk Helmrich (Jülich Supercomputing Centre)
    • 12:00 PM
      Lunch Break
    • 2
      PixelStreaming with Unreal Engine

      Introduction into WebRTC concepts, connectivity, and HPC usability

      Speaker: Dirk Helmrich (Jülich Supercomputing Centre)
    • 3
      WebRTC for Computer Vision Pipelines

      Encoding properties, setting up communication, including simulations, using data streams

      Speaker: Dirk Helmrich (Jülich Supercomputing Centre)
    • 12:00 PM
      Lunch Break
    • 4
      Building ML/AI Pipelines with Unreal Engine

      Preparing the frameworks
      Parsing and using data
      Best practices

      Speaker: Dirk Helmrich (Jülich Supercomputing Centre)