# Retargeting Motions to Diverse Skeletons via Learnable Flattening

arXiv: https://arxiv.org/abs/2609.38578 · HTML: https://mlsanity.com/paper/2609.38578/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 22 of its 39 references that are in the map.
- First posted on arXiv 29 Sep 2026.

## Builds on (most cited first, 22 of 22)

- [Graph Attention Networks](https://mlsanity.com/paper/1710.10903/) (2017)
- [Inductive Representation Learning on Large Graphs](https://mlsanity.com/paper/1706.02216/) (2017)
- [Expressive Body Capture: 3D Hands, Face, and Body from a Single Image](https://mlsanity.com/paper/1904.05866/) (2019)
- [AMASS: Archive of Motion Capture as Surface Shapes](https://mlsanity.com/paper/1904.03278/) (2019)
- [On the Continuity of Rotation Representations in Neural Networks](https://mlsanity.com/paper/1812.07035/) (2018)
- [Human Motion Diffusion Model](https://mlsanity.com/paper/2209.14916/) (2022)
- [Action-Conditioned 3D Human Motion Synthesis with Transformer VAE](https://mlsanity.com/paper/2104.05670/) (2021)
- [3D human pose estimation in video with temporal convolutions and semi-supervised training](https://mlsanity.com/paper/1811.11742/) (2018)
- [PhysDiff: Physics-Guided Human Motion Diffusion Model](https://mlsanity.com/paper/2212.02500/) (2022)
- [HuMoR: 3D Human Motion Model for Robust Pose Estimation](https://mlsanity.com/paper/2105.04668/) (2021)
- [Robust Motion In-betweening](https://mlsanity.com/paper/2102.04942/) (2021)
- [TEACH: Temporal Action Composition for 3D Humans](https://mlsanity.com/paper/2209.04066/) (2022)
- [Skeleton-Aware Networks for Deep Motion Retargeting](https://mlsanity.com/paper/2005.05732/) (2020)
- [Neural Kinematic Networks for Unsupervised Motion Retargetting](https://mlsanity.com/paper/1804.05653/) (2018)
- [Real-Time Style Modelling of Human Locomotion via Feature-Wise Transformations and Local Motion Phases](https://mlsanity.com/paper/2201.04439/) (2022)
- [Single Motion Diffusion](https://mlsanity.com/paper/2302.05905/) (2023)
- [Motion Puzzle: Arbitrary Motion Style Transfer by Body Part](https://mlsanity.com/paper/2202.05274/) (2022)
- [Learning Character-Agnostic Motion for Motion Retargeting in 2D](https://mlsanity.com/paper/1905.01680/) (2019)
- [AnyTop: Character Animation Diffusion with Any Topology](https://mlsanity.com/paper/2502.17327/) (2025)
- [Motion2Motion: Cross-topology Motion Transfer with Sparse Correspondence](https://mlsanity.com/paper/2508.13139/) (2025)
- [Learning Variational Motion Prior for Video-based Motion Capture](https://mlsanity.com/paper/2210.15134/) (2022)
- [PALUM: Part-based Attention Learning for Unified Motion Retargeting](https://mlsanity.com/paper/2601.07272/) (2026)

## Abstract (arXiv)

Cross-structural motion retargeting aims to transfer motion between different skeletal topologies. Despite recent progress, existing state-of-the-art models struggle with reliability in zero-shot settings, i.e. skeletons with different topologies which were unseen during training, and recent Transformer-based attempts have failed to outperform specialized geometric methods. We bridge this gap with a Transformer Autoencoder that learns a topology- and translation-invariant latent space. Our core contribution is a learnable flattening of skeletal graphs that captures both local dependencies and global structure. Unlike the standard transformer architecture, which adds positional information to token content, we integrate graph-based positional encodings multiplicatively, a design choice that follows directly from our flattening formulation. The resulting model handles diverse skeletal topologies within a single unified architecture and trains in a fully unsupervised manner, requiring no paired retargeting data. Ablation studies show, that the graph encodings, multiplicative formulation, and Transformer backbone is critical for the performance. In zero-shot evaluations, our method reduces global joint position error by 43-47% over current benchmarks. A user study (n = 37), including expert animators, further ranks our approach highest in motion alignment and physical plausibility (p < 0.05). These results demonstrate that our model design is key to making transformer architectures effective for motion retargeting, outperforming existing approaches.