← ArchivePaper2021

Direct Delta Mush Skinning Compression with Continuous Examples

Binh Huy Le, Keven Villeneuve, Carlos Gonzalez-Ochoa

SIGGRAPHEpicSkinning

Compresses DDM models into a two-layer LBS representation using continuous pose-space example data to minimize approximation error.

Open the paper ↗Trace in the lineage graph

Abstract

Direct Delta Mush (DDM) is a high-quality, direct skinning method with a low setup cost. However, its storage and run-time computing cost are relatively high for two reasons: its skinning weights are 4 X 4 matrices instead of scalars like other direct skinning methods, and its computation requires one 3 X 3 Singular Value Decomposition per vertex. In this paper, we introduce a compression method that takes a DDM model and splits it into two layers: the first layer is a smaller DDM model that computes a set of virtual bone transformations and the second layer is a Linear Blend Skinning model that computes per-vertex transformations from the output of the first layer. The two-layer model can approximate the deformation of the original DDM model with significantly lower costs. Our main contribution is a novel problem formulation for the DDM compression based on a continuous example-based technique, in which we minimize the compression error on an uncountable set of example poses. This formulation provides an elegant metric for the compression error and simplifies the problem to the common linear matrix factorization. Our formulation also takes into account the skeleton hierarchy of the model, the bind pose, and the range of motions.

How to read this

Category: Method: skinning compression for Direct Delta Mush
Contributions: A two-layer compression of a DDM model: a smaller DDM layer computing virtual bone transformations, plus a Linear Blend Skinning layer for per-vertex transforms, cutting storage and run-time cost
A novel continuous example-based problem formulation that minimizes compression error over an uncountable set of example poses
Reduction of the problem to standard linear matrix factorization while accounting for the skeleton hierarchy
Context: Compresses the Direct Delta Mush skinning model (and variants), replacing its per-vertex 4x4 weights and per-vertex SVD with a cheaper LBS-backed approximation.Builds on: Direct Delta Mush Skinning and Variants
Correctness: Aims to approximate the original DDM deformation at significantly lower cost; it is an approximation, so a reader should weigh the accuracy/cost trade-off and check how the virtual-bone count affects fidelity for their meshes.
Clarity: Fairly technical; a first pass gives the two-layer idea, a second pass is needed for the continuous-example formulation.
How to read it: First pass for the two-layer DDM-to-LBS structure and why it is cheaper; second pass on the continuous example formulation and matrix factorization if you implement or tune the compression.

Builds on

Direct Delta Mush Skinning and Variants 2019

Built upon by

Nothing yet.

Related work

Direct Delta Mush Skinning and Variants 2019 / SIGGRAPH
Two-Layer Sparse Compression of Dense-Weight Blend Skinning 2013 / SIGGRAPH
Fast Automatic Skinning Transformations 2012 / SIGGRAPH
Bodyopt: A Character Deformation Pipeline for Avatar: The Way of Water 2023 / SIGGRAPH

Keywords

direct delta mush skinning compression linear blend skinning virtual bones multi-weights two-layer model continuous example-based technique matrix factorization sparse coding dictionary learning singular value decomposition skeletal animation range of motion floating-point cancellation skin sliding least squares k-means clustering deformation

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 →