Multi-modal system for human interaction capture

This project will realise a motion capture (mocap) system for some of the most challenging genres of human motion: dexterous manipulation and grasping. Since tracking using optical sensors and cameras is often problematic due to occlusion, the proposed system will incorporate a custom ensemble of sensors using different modalities, such as optical markers, RGBD cameras, IMUs, and pressure sensors. Data from individual sensors will be synchronized and resampled in order to compile a high-quality database of human grasping and manipulation. This dataset will later be used to reconstruct dexterous manipulation and grasping motions for real-time motion synthesis in computer graphics and animation applications. 

Connaissances requises

Programming digital hardware (Raspberry Pi, Arduino, or similar)

  •  Python and/or C++ programming languages
  •  Strong mathematical skills (linear algebra, 3D math, geometry)
  •  Programming computer graphics applications or video game engines  

Programme d'études visé

Maîtrise avec projet, Maîtrise avec mémoire

Domaines de recherche

Technologies de l'information et des communications

Financement

Funding is available / Une bourse est disponible

Autres informations

Date de début : 2019-05-01

Personne à contacter

Sheldon Andrews | sheldon.andrews@etsmtl.ca