Student Projects and Competences
Magnetic-Inertial Sensor-Network
Sampling the shape of a human body or parts thereof is feasibly via motion capture or visual 3D scanning. Wearable techniques would allow physiotherapy and orthopedy researchers to observe people's e.g. back shape during everyday situations. Until today sensors with sufficient accuracy are not available. This student research project, proposes a sensor network of magnetic inertial sensor nodes, that have the potential to be integrated into a wearable textile.
Students: Hannes Scherrer & Sebastian Frey, BA FS2024 Jannis Gull, PA HS2025
Contact person: Prof. Dr. Martin Weisenhorn
Synthetic Dataset Generation for HPE Estimation
A synthetic data generation pipeline has been created using motion capture and Unreal Engine 5. The system produces realistic videos of digital athletes performing recorded exercises, complete with precise 2D and 3D joint annotations. By randomizing actors, environments, and camera settings, we achieve diverse datasets for robust AI training. This technology enables improved human pose estimation models for sports and health applications and lays the groundwork for intelligent motion analysis and feedback systems on consumer devices.
Students: Milena Squindo, PA FS 2024, Kai Erdin, SA HS 2024
Contact person: Prof. Dr. Martin Weisenhorn