Cages and Skeletons in Digital Animation: A Novel Skeleton-based Approach for Cage Generation

Cages together with skeletons are the most widely used structures to pose a digital character, giving the illusion of its movement. While skeletons are generally used for primary motion, like limbs movements, cages instead, are usually used to enrich primary motion with secondary effects, like body jiggling, character breath, cloth movements etc. In the light of the importance of cages in the animation pipeline, we have investigated two main subproblems related to cage-based deformations techniques: build the cage; provide powerful tools to perform assist in the process of deforming digital models. Cages are intrinsically more complicated to be created than skeletons. During their design, well-established properties have to be fulfilled: the cage must tightly envelop the original model without intersecting it; it must be coarse enough to be easily manipulated, and it must be shape-aware, i.e., its control nodes should be close to the parts of the model one would like to deform or bend. Due to these hard constraints, cages are often hand-made, and their creation may require hours of extensive work by skilled artists. For this reason, we focused on defining a novel skeleton based approach which allows the user to quickly design high-quality cages for animation; and developing a novel research-oriented software tool to perform cage-based deformations in a lightweight and easy to use environment. In this thesis after seeing the properties of the cages and the cage-based deformations, we will see the state-of-the-art of the existing cage generation method and their drawbacks. In the core of the thesis, we will look at the details of our approach, the results obtained and a comparison with the cages on state of the art. In conclusion, we will see our novel research-oriented software tool for the visualization, editing and generation of cage-based animation whose goal is to support the growing interest of the research community.