← ArchivePaper2008
Reusable Skinning Templates Using Cage-based Deformations
Tao Ju, Qian-Yi Zhou, Michiel van de Panne, Danny Cohen-Or, Ulrich Neumann
Introduces reusable per-joint cage deformation templates that transfer skinning across characters with similar skeletons.
Abstract
This paper introduces skinning templates, which define common skeleton-driven deformation behaviors for common joint and bone types so that skinning solutions can be shared and reused across different characters. The templates are implemented using cage-based deformations, where the skeleton drives the cage vertices through example-based deformation functions that preserve radial distance to avoid linear-blend artifacts such as pinching and collapse, and the cage in turn smoothly deforms the embedded geometry via Positive Mean Value Coordinates. A semi-automatic fitting procedure adapts templates to varying geometries while preserving deformation intent, enabling effects like muscle bulging and pinch-free elbow bending. Templates can be swapped interactively during an animation preview cycle, allowing rapid exploration of alternate skinning styles.
How to read this
- Category
- Method: a reusable cage-based skinning technique
- Contributions
- Introduces skinning templates that define skeleton-driven deformation behaviors per joint and bone type for reuse across characters
- Implements templates with cage-based, example-driven deformations that preserve radial distance to avoid linear-blend pinching and collapse, propagated to geometry via Positive Mean Value Coordinates
- Provides a semi-automatic fitting procedure and interactive template swapping for rapid skinning-style exploration
- Context
- Builds on Mean Value Coordinates for Closed Triangular Meshes (Ju et al. 2005) and addresses the classic linear-blend skinning artifacts by transferring deformation intent rather than weights.Builds on: Mean Value Coordinates for Closed Triangular Meshes
- Correctness
- Validated through demonstrated effects such as muscle bulging and pinch-free elbow bending; key assumptions are that target characters share similar skeleton structure and that templates capture intent, so behavior on dissimilar topologies or extreme poses warrants caution.
- Clarity
- Accessible; a first pass conveys the template-and-cage idea, do a second pass for the radial-distance deformation functions and the fitting procedure.
- How to read it
- Focus first on what a template is and why cages plus Positive MVC avoid pinching; a second pass on the example-based deformation functions and fitting is worth it if you plan to implement reuse.
Builds on
Built upon by
Related work
- Harmonic Coordinates for Character Articulation 2007 / SIGGRAPH
- Geodesic Voxel Binding for Production Character Meshes 2013 / SCA
- Automatic Rigging and Animation of 3D Characters 2007 / SIGGRAPH
- Stretchable and Twistable Bones for Skeletal Shape Deformation 2011 / SIGGRAPH Asia
Keywords
This page summarises the entry and links to its original source. The archive never hosts or redistributes the publication itself.Show it in the full archive list →