MoReFlow: Motion Retargeting Learning through Unsupervised Flow Matching

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kim, Wontaek, Li, Tianyu, Ha, Sehoon
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911218129698816
author Kim, Wontaek
Li, Tianyu
Ha, Sehoon
author_facet Kim, Wontaek
Li, Tianyu
Ha, Sehoon
contents Motion retargeting holds a premise of offering a larger set of motion data for characters and robots with different morphologies. Many prior works have approached this problem via either handcrafted constraints or paired motion datasets, limiting their applicability to humanoid characters or narrow behaviors such as locomotion. Moreover, they often assume a fixed notion of retargeting, overlooking domain-specific objectives like style preservation in animation or task-space alignment in robotics. In this work, we propose MoReFlow, Motion Retargeting via Flow Matching, an unsupervised framework that learns correspondences between characters' motion embedding spaces. Our method consists of two stages. First, we train tokenized motion embeddings for each character using a VQ-VAE, yielding compact latent representations. Then, we employ flow matching with conditional coupling to align the latent spaces across characters, which simultaneously learns conditioned and unconditioned matching to achieve robust but flexible retargeting. Once trained, MoReFlow enables flexible and reversible retargeting without requiring paired data. Experiments demonstrate that MoReFlow produces high-quality motions across diverse characters and tasks, offering improved controllability, generalization, and motion realism compared to the baselines.
format Preprint
id arxiv_https___arxiv_org_abs_2509_25600
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MoReFlow: Motion Retargeting Learning through Unsupervised Flow Matching
Kim, Wontaek
Li, Tianyu
Ha, Sehoon
Graphics
Robotics
Motion retargeting holds a premise of offering a larger set of motion data for characters and robots with different morphologies. Many prior works have approached this problem via either handcrafted constraints or paired motion datasets, limiting their applicability to humanoid characters or narrow behaviors such as locomotion. Moreover, they often assume a fixed notion of retargeting, overlooking domain-specific objectives like style preservation in animation or task-space alignment in robotics. In this work, we propose MoReFlow, Motion Retargeting via Flow Matching, an unsupervised framework that learns correspondences between characters' motion embedding spaces. Our method consists of two stages. First, we train tokenized motion embeddings for each character using a VQ-VAE, yielding compact latent representations. Then, we employ flow matching with conditional coupling to align the latent spaces across characters, which simultaneously learns conditioned and unconditioned matching to achieve robust but flexible retargeting. Once trained, MoReFlow enables flexible and reversible retargeting without requiring paired data. Experiments demonstrate that MoReFlow produces high-quality motions across diverse characters and tasks, offering improved controllability, generalization, and motion realism compared to the baselines.
title MoReFlow: Motion Retargeting Learning through Unsupervised Flow Matching
topic Graphics
Robotics
url https://arxiv.org/abs/2509.25600