Extending a shape-driven map to the interior of the input shape and to the surrounding volume is a difficult problem since it typically relies on the integration of shape-based and volumetric information, together with smoothness conditions, interpolating constraints, preservation of feature values at both a local and global level. In this context, this course revises the main out-of-sample approximation schemes for both 3D shapes and d-dimensional data, and provides a unified discussion on the integration of surface-and volume-based shape information. Then, it describes the application of shape-based and volumetric techniques to shape modeling and analysis through the definition of volumetric shape descriptors; shape processing through volumetric parameterization and polycube splines; feature-driven approximation through kernels and radial basis functions. We also discuss the Hamilton's Ricci flow, which is a powerful tool to compute the conformal structure of the shapes and to design Riemannian metrics of manifolds by prescribed curvatures and shape descriptors using conformal welding. We conclude the presentation by discussing applications to shape analysis and medicine, open problems, and future perspectives.
Course descriptionShape modeling typically handles a 3D shape as a two-dimensional surface, which describes the shape boundary and is represented as a triangular mesh or a point cloud. However, in several applications a volumetric surface representation is more suited to handle the complexity of the input shape. For instance, volumetric representations are used to accurately modeling the behavior of non-rigid deformations and volume constraints are imposed to avoid deformation artifacts. In shape matching, volumetric descriptors, such as Laplacian eigenfunctions, heat kernels, and diffusion distances, are defined starting from their surface-based counterparts.In the aforementioned applications, the underlying problem generally requires the prolongation of the surface-based information, which is typically represented as a shape-driven map, to the interior of the input shape or, more generally, to the surrounding volume. Extending a surface-based scalar function to a volumetric representation is a difficult problem since it typically relies on the integration of shape-based and volumetric information, together with smoothness conditions, interpolating constraints, preservation of feature values at both a local and global level. Besides the underlying complexity and degrees of freedom in the definition of volumetric approximations of surface-based maps, out-of-sample approximations (i.e., the extension of a surface-based scalar function to a volume-based approximation) are essential to address a wide range of problems. For instance, volumetric Laplacian eigenfunctions are suited to define volumetric shape descriptors, which are consistent with their surface-based counterparts. In a similar way, harmonic volumetric functions have been used for the volumetric parameterization and the definition of (volumetric...