From the video, it appears what Disney is doing is much simpler compared to the techniques in the article. It seems children are simply "filling in" parts of ex

From the video, it appears what Disney is doing is much simpler compared to the techniques in the article. It seems children are simply "filling in" parts of existing figures (e.g. the Mickey Mouse ears shown in the video), which are then animated. This seems to me like a less challenging problem, as all the rigging/joint/etc stuff that Meta is doing can be done by-hand, beforehand, in the Disney case, it does not need any ML. The cool thing with Meta's system is that it (hypothetically) works on arbitrary human-like figures.