Scene with skeleton superimposed

For each person in the scene we have generated two sets of virtual joint positions. They are derived from the 37 marker positions. In one set we use a form of averaging. In the other set, they are computed by means of kinematic contraints. For more details we refer to the HICV paper.

Virtual joint names

There are 15 virtual joint positions per person:

In the C3D files, a label consists of a person part and a marker part, separated by a colon. E.g., N:lshoulder is the left shoulder virtual joint position of person N.