Models for deformable objects modeling thesis proposal

The evolution of Computer Graphics methodologies in the last two decades has been impressive, opening many opportunities to work on new problems and applications. Sophisticated techniques are available for visualizing complex scenes and to support user interaction and im­mersion with virtual worlds. An active research domain ls nowadays how to extend graphics technology to include in interactive and immersive graphic contexts those objects whose shape could be deformed depending on the forces that act on them. The problem becomes more difficult if we require that these deformations have to be physically based, i.e. coherent with some properties which the objects are supposed to have. The field of Soft Object Modeling or, more generally, Soft Object Animation has been investigated for the last ten years and a wide range of very different approaches has been proposed, with a main orientation to animation and offline simulation. Following this trend, we intend to introduce concepts and techniques strictly related to interactive graphics modeling and visualization in order to develop methods and data representations for Soft Object Modeling that could be used for immersive and interactive applications. In this field, the classical compromise is between real-time performance and accuracy. Our aim is not to search methodology that could improve simulation accuracy (in the sense of more sophisticated descriptions of deformation), but to search for feasible techniques that should produce realistic simulations with a much cheaper processing cost, to fulfill the constraint of interactive/real time feedback. The application domain where the methodologies proposed will be evaluated is the virtual simulation of soft tissue for medical applications. This choice will introduce a number of application-oriented constraints, but we plan to design highly general representation techniques and methodologies that should not be restricted to the test application domain.