Interactive motion decomposition

Figure 1: In the left image the original walk (left actor) has been modified with a ”sneaky” style, yielding a sneaky walk (right actor). In the right image, the sneaking motion (left actor) is modified with a ”walking” style, yielding a walk-like sneak that appears as a tiptoeing motion (right actor). Our visual decomposition works with any sort of motion data, including motion captured and keyframed data.

[1]  Aaron Hertzmann,et al.  Style machines , 2000, SIGGRAPH 2000.

[2]  Frédéric H. Pighin,et al.  Unsupervised learning for speech motion editing , 2003, SCA '03.