What it is about
You learn methods for 3D capture and motion capture with a focus on RGB-D sensors and algorithmic fundamentals. In the end you can capture data with consumer RGB-D scanners, understand rigid and non-rigid 3D reconstructions, pose and face tracking, as well as suitable optimization methods, and you can practically implement selected approaches.
What you will be able to do
- Understanding rigid and non-rigid 3D reconstruction
- Applying RGB-D scanning with devices like Kinect, Tango, RealSense
- Use of ICP and camera tracking methods
- Knowledge of sensor calibration and volumetric fusion
- Familiarity with pose, body, facial and hand tracking
- Foundations of 3D deep learning and relevant optimization methods (GN, LM, Gradient Descent)
- Practical implementation and application of the methods in projects
What the module consists of
- VorlesungConveys subject knowledge, shows examples and refers to relevant articles
- Übungen / HomeworksDeepens lecture content through practical tasks
- Projekt (ca. 2 Monate)Includes a small research/implementation project to apply the learned methods
Teaching method
- Vorlesung mit DemonstrationenExplains concepts and demonstrates applications on concrete examples
- Literaturhinweise und EigenstudiumEncourages reading relevant articles and comparing approaches
- Übungsaufgaben als HausaufgabenEnable practical implementation and consolidation of the material
- AbschlussprojektLets you work on a topic practically and implement it independently