My first thought is to talk about things that have to do with motion capture and film. Motion capture technology is getting better and is being used more and more in animation and games. We should think about whether we should give up on traditional animation and focus on motion capture instead.
In the past few years, people have become more interested in motion capture systems. On the other hand, current systems only have a few degrees of freedom, so they can only get the motion of rigid bodies that are linked together. We show a system for capturing surfaces that change shape, like moving cloth, that includes both geometry and parameterization. We figure out the shape of the cloth using stereo correspondence, and we use the scale-invariant feature transform (SIFT) to find any pattern printed on it, even when the cloth is moving quickly. We explain a new seed-and-grow method for adapting the SIFT algorithm to geometry that can change shape. Lastly, we parameterize the whole geometry by interpolating feature points.
Motion capture is different from making animation by watching people move. For one thing, motion capture can be used for things other than animation, like biomedical analysis, surveillance, sports performance analysis, or as a way for people and computers to talk to each other. Each of these tasks is both different and the same as the problems that come up when making animation. At the first step of each, you must make the observations that will be used to figure out what’s going on, e.g., record the movements. Many of the ways animation is done come from the fields of biomechanics or medicine.