EventCap: Monocular 3D Capture of High-Speed Human Motions using an Event Camera

Background: 3D Human Motion Capture

Multi-view infra-based method → Multi-view markerless method → Single-view (monocular) with depth camera method → Multiple event camera method → Monocular event camera method (with help of deep learning)

Infra-based motion capture

https://d10lvax23vl53t.cloudfront.net/images/Article_Images/ImageForArticle_43_15816813570943783.png

https://www.researchgate.net/profile/Francis-Quek/publication/220913524/figure/fig2/AS:305583229030401@1449868127306/Data-collection-an-optical-motion-capture-system-and-a-3D-acceleration-sensor-based-data.png

Limitation of infra-based motion capture

Markerless image-based method

Overcome the limitation of infra-based motion capture method

Multi-Camera system

https://www.researchgate.net/profile/Peter-With-2/publication/220785617/figure/fig2/AS:667707607379978@1536205307599/Multiview-capturing-system-in-which-the-position-and-orientation-of-each-camera-is-known.png

Limitation of multi-camera method (including monocular depth method)