TOG
2008
13 years 11 months ago
2008
This paper proposes a new marker-less approach to capturing human performances from multi-view video. Our algorithm can jointly reconstruct spatio-temporally coherent geometry, mo...