← ArchivePaper2021
Contact-Aware Retargeting of Skinned Motion
Ruben Villegas, Duygu Ceylan, Aaron Hertzmann, Jimei Yang, Jun Saito
A geometry-conditioned recurrent network with encoder-space optimization preserves self-contacts and prevents interpenetration when retargeting motion across different character bodies.
Abstract
This paper introduces a motion retargeting method that preserves self-contacts and prevents interpenetration. Self-contacts, such as when hands touch each other or the torso or the head, are important attributes of human body language and dynamics, yet existing methods do not model or preserve these contacts. Likewise, interpenetration, such as a hand passing into the torso, are a typical artifact of motion estimation methods. The input to our method is a human motion sequence and a target skeleton and character geometry. The method identifies self-contacts and ground contacts in the input motion, and optimizes the motion to apply to the output skeleton, while preserving these contacts and reducing interpenetration. We introduce a novel geometry-conditioned recurrent network with an encoder-space optimization strategy that achieves efficient retargeting while satisfying contact constraints. In experiments, our results quantitatively outperform previous methods and we conduct a user study where our retargeted motions are rated as higher-quality than those produced by recent works. We also show our method generalizes to motion estimated from human videos where we improve over previous works that produce noticeable interpenetration.
How to read this
- Category
- Method: contact-aware motion retargeting across skinned characters
- Contributions
- A motion-retargeting method that detects self-contacts and ground contacts in the input and preserves them while reducing interpenetration on the target body
- A geometry-conditioned recurrent network with an encoder-space optimization strategy to satisfy contact constraints efficiently
- Generalization to motion estimated from human video, improving over prior estimation
- Context
- Builds on deep motion retargeting (e.g. Skeleton-Aware Networks), adding character geometry conditioning and explicit self-contact and interpenetration handling.Builds on: Skeleton-Aware Networks for Deep Motion Retargeting
- Correctness
- Reported to quantitatively outperform prior methods with a user study rating its retargeted motion higher; note that results hinge on reliable contact detection in the input, and the method targets skinned-character retargeting rather than full physical plausibility.
- Clarity
- Accessible at a first pass for the problem and approach; a second pass clarifies the encoder-space optimization.
- How to read it
- First pass for the contact-preservation framing and why geometry conditioning matters; second pass on the network plus encoder-space optimization if you build retargeting pipelines.
Builds on
Related work
- Geometry-Aware Retargeting for Two-Skinned Characters Interaction 2024 / SIGGRAPH Asia
- Learning Character-Agnostic Motion for Motion Retargeting in 2D 2019 / SIGGRAPH
- Normalized Euclidean Distance Matrices for Human Motion Retargeting 2017 / MIG
- Dog Code: Human to Quadruped Embodiment Using Shared Codebooks 2024 / MIG
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 →