Computer Graphics Recipes for Endowing 3D Shapes with Semantics in Virtual Worlds

The massive use of 3D models increased also in the Gaming community the awareness of the benefits of a digital representation of objects involving not only geometric data but also their semantics. Encoding semantic information (that is, functionality or meaning) gives the opportunity of performing automatic reasoning on object's features, and interaction capabilities, behaviour and constraints. In a virtual environment scenario, associating 3D shapes with semantics may allow for smart search and retrieval of characters and objects, fast re-use of animation sequences and easy re-adaptation of content to new tasks and environments. Nonetheless, the process of semantic annotation should require the lowest effort from the designer side to be really feasible and applicable. In this paper we are going to present the recent research outcomes of Computer Graphics devoted to the segmentation of 3D shapes into main components and to the identification of salient features, to the extraction and representation of knowledge related to 3D content, and to the association of such knowledge with the geometry of objects (to their components or as a whole). These, in fact, can be seen as the basic lements to build semi-automatic tools to support the semantic annotation of 3D shapes.