Pro­ject Group: 3D-Ma­chi­ne-Lear­ning (3D-ML)

Supervisor: Tobias Nickchen

Description:

The project group "3D Machine Learning" deals with the analysis of 3D-Data in the context of Additive Manufacturing. The problem arose in cooperation with the 3D printing service provider Protiq, whereby there is a real connection to the industry. In the domain of additive manufacturing there are a number of problems for which a 3D-Model-Analysis is necessary. From the determination of production costs to technical analyses with regard to stresses, conductivity or other characteristic values up to the automatic sorting of components. The question considered here deals with a step which is located in the workflow of additive manufacturing after the actual manufacturing process: The automated object recognition of the produced parts.

At the Protiq GmbH and also at other 3D printing service providers most parts are produced by Selective Laser Sintering (SLS). With this production process, a large number of different components can be produced in a three-dimensional nested package in one construction space. The different parts belong to different customers and different orders. Therefore it is necessary to assign the various objects to the correct orders after production.To do this manually consumes a lot of time, which is why the possibility of automation should be discussed in this project group.

As part of the project group, the participants should first familiarize themselves with the domain of 3D-Machine-Learning and specifically Deep Learning. Currently, there are various approaches for the analysis of 3D-Models. These will first be analyzed theoretically. Based on the analysis, the most promising approaches will then be implemented, trained and evaluated based on test data sets.

To find promising solution, there are several questions to be answered:

  • What are the requirements for the solution?
  • How can the requirements be met?
  • How can we measure the achievement of our goals?

Requirements

  • (required) Theoretical knowledge in the field of Machine Learning
      - The successful conclusion of one or more courses with ML reference would be optimal
  • (required) Ability to read, understand and write C/C++/Python code
  • (nice-to-have) Practical experience in the field of Machine Learning

Course information

  • The project group will be conducted in English (unless all participants agree with the German language)
  • Computer science (CS) and engineering (CE) students are welcome
      - CS students get 20 ECTS, CE students 18 ECTS