Robust single-view geometry and motion reconstruction

Abstract

input scans reconstruction input scans reconstruction input scans reconstruction Figure 1: Reconstruction of complex deforming objects from high-resolution depth scans. Our method accurately captures the global topology and shape motion, as well as dynamic, small-scale details, such as wrinkles and folds. AbstractWe present a framework and algorithms for robust geometry and motion reconstruction of complex deforming shapes. Our method makes use of a smooth template that provides a crude approximation of the scanned object and serves as a geometric and topological prior for reconstruction. Large-scale motion of the acquired object is recovered using a novel space-time adaptive, non-rigid registration method. Fine-scale details such as wrinkles and folds are synthesized with an efficient linear mesh deformation algorithm. Subsequent spatial and temporal filtering of detail coefficients allows transfer of persistent geometric detail to regions not observed by the scanner. We show how this two-scale process allows faithful recovery of small-scale shape and motion features leading to a highquality reconstruction. We illustrate the robustness and generality of our algorithm on a variety of examples composed of different materials and exhibiting a large range of dynamic deformations.

Object Deformations From Image Silhouettes Using a Kinematic Finite Element Beam Model

“…Possible approaches are volumetric approaches which are widely used in matching of anatomical data or approaches based on the surface of an object. They use NURBS or mesh representations of a given model which is deformed by applying a deformation graph or function (Jordt & Koch, 2012;Li et al, 2009). With such a shape representation it is possible to determine the deformation of an object even out of only one image (Salzmann et al 2007).…”

“…In Li et al (2009) the deformation graph describes the deformation of a 3D object. By using the assumptions of an Euler-Bernoulli beam this graph could be reduced to 1D.…”

Non-rigid Registration Meets Surface Reconstruction

“…After applying these local deformations, the whole mesh is undergoes a global rigid transformation. This model has been improved in [19] after replacing the uniform sampling by a temporally adaptive distribution that refines the deformation graph. Articulated models are described in [9] through a set of rigid transformations that are associated to each point using skinning weights.…”

“…In [10] and [25] further geometric information including normals and curvatures are exploited for better distance estimations. The authors in [19] employed a combination of point-to-point and point-to-plane distances for developing a fitting energy term in non-rigid registration.…”

“…The outcome is a function in the form of non-linear least squares that is solved through the Levenberg-Marquadt algorithm. Li et al in [20,19] apply this framework for minimizing a comprehensive energy term for non-rigid registration. In [29], the optimal TPS parameters are estimated by solving a nonlinear system of equations that are obtained through computing some proper integrals over the mesh.…”