PriMoRetarget: Humanoid Motion Transfer Through Prioritized Motion Retargeting

Nicholas Ioannidis; Dun Yang; Max Smith; Tue Vu; Ziyu Zhang; Shayegan Omidshafiei; Xue Bin Peng

cs.ROOct 8, 2026

PriMoRetarget: Humanoid Motion Transfer Through Prioritized Motion Retargeting

Authors: Nicholas Ioannidis, Dun Yang, Max Smith, Tue Vu, Ziyu Zhang, Shayegan Omidshafiei, Xue Bin Peng

Organizations: Simon Fraser University · FieldAI · NVIDIA · Alberta Machine Intelligence Institute

Abstract

Motion retargeting is a critical preprocessing step for transferring human motion data to humanoid robots. While existing methods are effective for single-character motions, interactions involving multiple characters and objects introduce additional contacts and spatial relationships that must be preserved. Retargeting is commonly formulated as a weighted multi-objective optimization problem, where conflicting objectives can require careful tuning of relative weights and can still compromise important motion characteristics. We introduce PriMoRetarget, a prioritized motion-retargeting framework that replaces relative objective weighting with an explicit priority hierarchy. Lower-priority updates are constrained to preserve higher-priority objectives, allowing critical characteristics such as foot placement and interaction contacts to take precedence over lower-priority pose objectives. The method is evaluated on four datasets spanning single-character motion, character-object manipulation, and character-character interactions. The resulting motions better preserve task-relevant interactions and provide reference trajectories that can be tracked more reliably than those produced by widely used retargeting methods.

Figures & tables

Fig. 2: Weighted versus prioritized optimization for two competing objectives. The weighted solution compromises between objectives, while prioritized optimization restricts lower-priority updates to directions that preserve the optimal solution for Objective 1.

Pose tracking , for keypoint $i$
Objective	Error function $f$
Position	$\hat{p}_{t,i}-s_{\dagger}\cdot\bar{p}_{t,i}$
Rotation	$\hat{R}_{t,i}\ominus\bar{R}_{t,i}$
Interaction , for keypoint pair $(i,j)$
Self	$(\hat{p}_{t,i}-\hat{p}_{t,j})-s_{\mathrm{s}\rightarrow\mathrm{t}}(\bar{p}_{t,i}-\bar{p}_{t,j})$
Cross	$(\hat{p}^{a}_{t,i}-\hat{p}^{b}_{t,j})-s_{\dagger}(\bar{p}^{a}_{t,i}-\bar{p}^{b}_{t,j})$

TABLE I: Error functions used in retargeting objective.

		Penetration $\downarrow$		Sliding $\downarrow$	Floating $\downarrow$	Jitter $\downarrow$	Contact $\uparrow$	RL tracking
Interaction Type	Method	dur [%]	depth [cm]	vel [cm/s]	height [cm]	[m/s 3 ]	[%]	succ [%] $\uparrow$	err [cm] $\downarrow$	obj [cm] $\downarrow$
Single-character (LAFAN1)	GMR	$12.1^{\pm 19.1}$	$10.7^{\pm 8.6}$	$5.2^{\pm 1.8}$	$2.1^{\pm 0.5}$	$216.7^{\pm 71.4}$	N/A	$90.0^{\pm 30.0}$	$4.0^{\pm 6.1}$	N/A
	OmniRetarget	$\mathbf{1.5}^{\pm 3.8}$	$\mathbf{0.7}^{\pm 1.0}$	$9.4^{\pm 2.7}$	$\mathbf{0.6}^{\pm 0.5}$	$229.4^{\pm 64.0}$	N/A	$90.0^{\pm 30.0}$	$5.2^{\pm 11.3}$	N/A
	Ours	$12.2^{\pm 21.7}$	$3.9^{\pm 3.8}$	$\mathbf{3.5}^{\pm 1.1}$	$1.4^{\pm 0.2}$	$\mathbf{135.6}^{\pm 52.7}$	N/A	$\mathbf{100.0}^{\pm 0.0}$	$\mathbf{1.8}^{\pm 0.7}$	N/A
	Source	N/A	N/A	$1.2^{\pm 0.2}$	$0.4^{\pm 0.4}$	$126.5^{\pm 52.4}$	N/A	N/A	N/A	N/A
Character–object (OMOMO)	PHC	$94.1^{\pm 9.9}$	$9.3^{\pm 3.6}$	$12.9^{\pm 3.2}$	$\mathbf{0.2}^{\pm 0.3}$	$\mathbf{76.3}^{\pm 14.6}$	$63.2^{\pm 24.5}$	$86.6^{\pm 34.0}$	$7.4^{\pm 8.5}$	$31.1^{\pm 70.3}$

TABLE II: Kinematic motion quality and tracking performance via RL.

Fig. 3: Comparison of retargeting methods on a box-manipulation sequence. Baseline methods tend to distort the interactions or lose contact with the box, while PriMoRetarget preserves the interactions more closely.

Single-character (LAFAN1)
	Penetration $\downarrow$		Sliding $\downarrow$	Floating $\downarrow$	Jitter $\downarrow$	Contact $\uparrow$
Variant	dur [%]	depth [cm]	vel [cm/s]	height [cm]	[m/s 3 ]	[%]
Ours	$\mathbf{12.2}^{\pm 21.7}$	$\mathbf{3.9}^{\pm 3.8}$	$3.5^{\pm 1.1}$	$1.4^{\pm 0.2}$	$\mathbf{135.6}^{\pm 52.7}$	N/A
WS-uniform	$12.4^{\pm 21.8}$	$5.4^{\pm 5.0}$	$3.3^{\pm 0.9}$	$1.1^{\pm 0.2}$	$140.8^{\pm 54.9}$	N/A
WS- $10\times$	$13.1^{\pm 22.0}$	$6.1^{\pm 5.3}$	$\mathbf{2.9}^{\pm 0.8}$	$\mathbf{0.8}^{\pm 0.2}$	$141.4^{\pm 55.0}$	N/A
Character–object (OMOMO)

TABLE III: Comparison against weighted-sum variants

Fig. 4: Loss curves for three objectives with different priority levels on a representative motion clip. Prioritized optimization reduces the higher-priority foot and contact losses at the expense of the lower-priority body objective, reflecting the relative importance of the objectives in the hierarchy.

Single-character (LAFAN1)
	Penetration $\downarrow$		Sliding $\downarrow$	Floating $\downarrow$	Jitter $\downarrow$	Contact $\uparrow$
Variant	dur [%]	depth [cm]	vel [cm/s]	height [cm]	[m/s 3 ]	[%]
Ours	$\mathbf{12.2}^{\pm 21.7}$	$\mathbf{3.9}^{\pm 3.8}$	$3.5^{\pm 1.1}$	$1.4^{\pm 0.2}$	$\mathbf{135.6}^{\pm 52.7}$	N/A
No warmup	$90.1^{\pm 15.3}$	$14.1^{\pm 5.2}$	$10.2^{\pm 2.1}$	$\mathbf{0.0}^{\pm 0.1}$	$168.1^{\pm 59.5}$	N/A
Free-last	$17.3^{\pm 23.0}$	$6.9^{\pm 5.4}$	$\mathbf{2.7}^{\pm 0.6}$	$0.4^{\pm 0.2}$	$160.4^{\pm 56.1}$	N/A
Character–object (OMOMO)

TABLE IV: Ablation of the weighted warmup and free objectives.

Fig. 5: Runtime cost of forming the nullspace projection over different sequence lengths. The dense solve scales cubically, whereas our per-frame formulation remains nearly constant, largely dominated by GPU overhead.

Fig. 6: Hardware deployment of retargeted box-manipulation motions on the Unitree G1. Top: OMOMO manipulation sequence involving kicking and picking up a box. Bottom: Collaborative manipulation sequence collected in-house involving two robots picking up and placing down a large box.

Fig. 7: PriMoRetarget generalizes to different embodiments, scenes, and augmentations: a) a Unitree G1 and a Deep Robotics DR02 jointly carrying a box b) a Unitree H1-2 carrying a box down a ramp, and c) a box-carrying motion retargeted from the source (left) to a Booster T1 (center) and to a Unitree G1 (right), with the G1 holding a box enlarged by $1.5\times$ .

Motion Retargeting

Motion retargeting transfers movement across different bodies, skeletons, or robots while preserving its intent and physical feasibility.

See all 104 papers

OTRetarget: Joint Robot and Object Motion Retargeting via Optimal Transport
Sep 29, 2026Guillaume Besset, Erwann Carn, Timothée Carecchio +5Robotic Loco-Manipulation Contact-Rich Robotic Manipulation
Inria, Département d’Informatique de l’École Normale Supérieure, PSL Research University, Paris, France
ReFM: Semantic-Aware Refinement Flow Model for Motion Retargeting
Sep 25, 2026Jingxiang Qu, Lucie Taglienti, Evan AthertonMotion Retargeting
Autodesk Research
Human2Humanoid: Physics-Aware Cross-Morphology Motion Retargeting for Humanoid Robots
Jun 2, 2026Tianchen Huang, Feiyang Yuan, Junchi Gu +5Cross-Embodiment Robot Learning Motion Retargeting
Institute of Humanoid Robots, Department of Precision Machinery and Precision Instrumentation, University of Science and Technology of China, Hefei, Anhui 230026, China

Sep 29, 2026J/K move · Enter open · S save

OTRetarget: Joint Robot and Object Motion Retargeting via Optimal Transport

Guillaume Besset, Erwann Carn, Timothée Carecchio, Valentin Tordjman-Levavasseur+4

Inria, Département d’Informatique de l’École Normale Supérieure, PSL Research University, Paris, France

Robotic Loco-Manipulation Contact-Rich Robotic Manipulation Whole-Body Robotic Manipulation Motion Retargeting Human-to-Robot Motion Imitation

Open paper AskPDF arXiv

Transferring human motion to humanoid robots requires adapting the demonstrated motion to the robot morphology while preserving interactions with the environment. This is particularly challenging for loco-manipulation tasks, where contacts with the ground and manipulated objects must remain consistent despite differences in body proportions. Yet, skeletal motion alone does not fully describe these interactions, and fixing object trajectories limits the adaptation to a new embodiment. In this paper, we introduce OTR ETARGET, a unified approach to jointly retarget robot and multi-object motion from human demonstrations. Our approach represents surface interactions through signed distances, closest surface points, and relative directions, and uses entropic optimal transport to transfer these quantities across human, robot, and object geometries. We incorporate the resulting interaction targets into a constrained inverse kinematics formulation that balances contact preservation with motion style and jointly optimizes robot and object poses at each frame. This formulation accommodates robot-object and object-object interactions without rescaling the scene or the demonstration. We validate the proposed approach on OMOMO, where it achieves a robot- object interaction Jaccard score of 87% and a depth error of 8.7 mm, compared with 28% and 29.3 mm for OmniRetarget. Finally, we demonstrate transfer to a physical G1 humanoid using whole-body policies trained with reinforcement learning on the retargeted references, across motions including two-handed box pick-and-place onto a table.

Abstract

Motion Retargeting

Explore similar work

OTRetarget: Joint Robot and Object Motion Retargeting via Optimal Transport

ReFM: Semantic-Aware Refinement Flow Model for Motion Retargeting

Human2Humanoid: Physics-Aware Cross-Morphology Motion Retargeting for Humanoid Robots