Some of the most labor-intensive tasks in animation are the realistic modeling of bodies and faces as well as the simulation of motion. Computer Vision techniques, however, offer the means to derive the models faster and in a more realistic fashion.
In this talk, we will concentrate on video-based approaches because of their comparatively lower cost and better control of the dynamic nature of the processes. We will introduce some of the animation models that are commonly used and we will discuss:
The ability to model moving humans has applications ranging from entertainment to sports medicine and athletic training. Furthermore, although realism requires complex models that may be controlled by a large number of parameters, that number is currently limited by the ability of the graphics designer to set all of them. Automating the modeling process will help solve this problem and will allow an increase in realism while reducing the cost.