Joint-aware manipulation of deformable models

Abstract

Complex mesh models of man-made objects often consist of multiple components connected by various types of joints. We propose a joint-aware deformation framework that supports the direct manipulation of an arbitrary mix of rigid and deformable components. First we apply slippable motion analysis to automatically detect multiple types of joint constraints that are implicit in model geometry. For single-component geometry or models with disconnected components, we support user-defined virtual joints. Then we integrate manipulation handle constraints, multiple components, joint constraints, joint limits, and deformation energies into a single volumetric-cell-based space deformation problem. An iterative, parallelized Gauss-Newton solver is used to solve the resulting nonlinear optimization. Interactive deformable manipulation is demonstrated on a variety of geometric models while automatically respecting their multi-component nature and the natural behavior of their joints.

Meta-representation of shape families

“…Realizing this goal amounts to detecting relationships between shape parts or features, and using these relationships to constrain the shapes during editing, referred to as the analyzeand-edit approach. For example, Xu et al [2009] use slippable motion analysis to detect joints on the shapes, which can then be used as articulations to deform the shapes; Gal et al [2009] detect feature curves (wires) on the objects, whose spatial relationships are preserved when deforming the shape; Li et al [2010] introduce arterial snakes as feature curves used to deform 3D shapes that are inherently 1D, while Zheng et al [2011] propose instead to use spatial controllers (similar to cages) placed on shape parts as the primitives for editing. In a different context, Lin et al [2011] propose an analyze-and-edit approach for retargeting of 3D architecture, where the input models are analyzed and decomposed into a set of 1D structures that are easier to retarget.…”

“…As part of this effort, single shapes, particularly man-made objects, have been analyzed to extract semantic relations that can be made useful in various applications. One example is in applying constraints on a shape editing process, thereby achieving an intelligent edit where the prescribed geometric characteristics of the manipulated shape are preserved [Gal et al 2009;Xu et al 2009;Li et al 2010;Zheng et al 2011]. Ideally, the aim is to understand the essence of the family of shapes directly from their object geometry, in order to use this knowledge in the applications to determine the geometric shape and configuration of shape parts.…”

Recovering Functional Mechanical Assemblies from Raw Scans

“…Our work is also inspired by recent works on functionality recognition and analysis [11], [12], [13], [14], [15]. Xu et al [11] performed slippage analysis over contact surfaces to segment and categorize joints in man-made objects, and used the information for interactive volumetric-based space deformation.…”

“…Xu et al [11] performed slippage analysis over contact surfaces to segment and categorize joints in man-made objects, and used the information for interactive volumetric-based space deformation. Guo et al [12] analyzed the individual parts using sharp edge loops and extracted the contact faces between each pair of neighboring parts and then clustered the sets of the individual parts into meaningful sub-assemblies used for a hierarchical decomposition.…”

Illustrating the disassembly of 3D models

“…Recently, understanding the higher-level representations of 3D models has received considerable interest (see survey [11]), including symmetry detection [12,13,14], upright orientation inference [15] and shape abstractions [16]. Benefitting from these techniques, a wide spectrum of applications are spawned, such as shape manipulation [17,18,19], shape synthesis [20], motion analysis [21,22,23,24] and computational furniture design [25]. In addition, many works focus on the analysis of individual parts within an input model, which are similar to our framework.…”