So the explainer video is, and paper

The idea is to have the 3D model carry its 3D Silhouette Ribbons. So the artist’s 2D drawing can be mapped to the 3D Silhouette Ribbons. When the model moves, the 3D Silhouette Ribbons follows and carries the corresponding 2D drawing.

The process seems simplified by 2D. First data of 3D Silhouette Ribbons and motion field in 2D screen space are precomputed. Then artist draws lines, computer maps it to 3D Silhouette Ribbons. Motion is directed by the 2D motion field in screen space. The 3D Silhouette Ribbons mapped to screen is moved by the 2D motion field. During a move, the 2D drawing of its 3D Silhouette Ribbon is moved by <> (don’t understand :_( ). The artist is responsible to draw new lines for hidden 3D Silhouette Ribbons appears into the screen.


Figure out this thing:

Screen Shot 2020-01-16 at 8.35.01 PM.png