The main goal of this thesis will be to design and implement an online system for body stretching assistance. The system will have the form of a game where online users can do stretching exercises, collect points and compete with other players. Pose estimation from video stream will be used to detect body position and provide visual suggestions how to correct the pose to perform stretching correctly. 3D virtual character overlay, visual effects, challenges, levels, multiplayer competition or other gamification aspects may be included. The system will be deployed as a mobile application. The thesis will be part of an international research project.


  • Knowledge of English language (source code comments and final report should be in English)
  • Knowledge of web based APIs, is advantageous
  • Knowledge of body tracking systems like OpenPose is advantageous


The project should be implemented using a pose estimation methods based on TensorFlow lite library . Server-client architecture will be utilized to implement multi-user gamification aspect. Game engine like Unity3D might be also used.


For more information please contact Peter Kán, Tiare Feuchtner, Hannes Kaufmann.



Master Thesis