Superquadric based hierarchical reconstruction for virtualizing free form objects from 3D data

Superquadric based (SQ-based) model is used as the universal object description in AVR (from actual reality to virtual reality). However, due to the intrinsic property of superquadrics, its models are not flexible enough to characterize precisely the objects with asymmetry and free form surface. A hierarchical reconstruction approach for virtualizing these objects from 3D data in AVR is developed. Firstly, an initial approximation is produced by a superquadric fit to the 3D data. Then, the crude superquadric fit is refined by fitting the residue with global and local direct manipulation of free-form deformation (DMFFD). The key elements of the hierarchical method, including superquadric fit to 3D data, computation of distance maps, recursive-fitting algorithm for DMFFD, adaptive refinement and decimation of polygon mesh under DMFFD are proposed. The proposed approach is shown competent and efficient for virtualizing the complex objects into a virtual environment