Computer Vision Engineer
Evaluate and benchmark markerless pose estimation models in realistic home-therapy environments. Optimise the full vision pipeline for extracting 3D skeletal data. Focus on enhancing robustness against lighting variation, occlusions, and clutter. Develop real-time machine learning models to semantically interpret patient movements and classify therapy exercises. Prototype real-time integration of pose estimation into the Kymu telerehabilitation platform that runs on mobile devices.
Biomedical Data Engineer
Develop a robust data pipeline to ingest, clean, and standardise raw kinematic data from markerless pose estimation. Validate data quality and processing steps to ensure reliability for downstream clinical and ML analysis. Explore data representations that support clinical decision-making in remote rehabilitation scenarios.
Game Designer
Design and build novel exergames using computer vision: translate therapeutic movements into intuitive gameplay mechanics. Validate prototypes through iterative user testing with patients and clinicians. Develop and test motivational strategies (game mechanics, rewards, narrative) that improve patient adherence in unsupervised, home-based rehabilitation settings.