pith. machine review for the scientific record. sign in

arxiv: 2604.07331 · v1 · submitted 2026-04-08 · 💻 cs.RO · cs.AI· cs.CV

Recognition: 2 theorem links

· Lean Theorem

RoSHI: A Versatile Robot-oriented Suit for Human Data In-the-Wild

Wenjing Margaret Mao , Jefferson Ng , Luyang Hu , Daniel Gehrig , Antonio Loquercio
Authors on Pith no claims yet

Pith reviewed 2026-05-10 17:42 UTC · model grok-4.3

classification 💻 cs.RO cs.AIcs.CV
keywords human motion capturewearable sensorsegocentric perceptionfull body pose estimationrobot learningIMU fusionSLAM
0
0 comments X

The pith

A hybrid wearable fuses sparse IMUs with egocentric cameras to estimate full 3D body pose and shape in a metric global frame.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces RoSHI as a portable system that records complete human body movements during fast, real-world actions by combining inexpensive motion sensors on the body with head-mounted cameras that track the environment. The goal is to gather the kind of long, natural interaction data needed to train robots without relying on fixed camera setups or post-processing in a lab. If the approach works, researchers could collect usable motion sequences from everyday settings that remain accurate even when parts of the body are hidden or moving quickly. The authors test this on agile activities and show the results support direct use in humanoid robot policy training.

Core claim

RoSHI fuses low-cost sparse IMUs with egocentric SLAM from the glasses to estimate the wearer's full 3D pose and body shape in a metric global coordinate frame, using the IMUs for robustness against occlusions and high-speed motion while the SLAM component anchors long-horizon motion and stabilizes upper-body estimates.

What carries the argument

The hybrid sensor fusion of sparse IMUs for occlusion- and speed-robustness with egocentric SLAM for long-horizon anchoring and upper-body stabilization.

If this is right

  • Human motion data can be collected portably during agile real-world tasks without studio equipment.
  • The resulting sequences outperform other egocentric-only baselines on the collected dataset.
  • Performance reaches levels comparable to an exocentric state-of-the-art method on the same agile activities.
  • The recorded motions can be used directly to train humanoid robot policies in realistic settings.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same fusion principle could be applied to longer daily activities where drift becomes the dominant error source.
  • Removing the need for external anchors might allow motion capture inside homes or vehicles for robot training.
  • Future versions could test whether adding minimal additional IMUs further reduces upper-body drift during extreme motions.

Load-bearing premise

The IMUs and egocentric cameras will continue to complement each other on fast, occluded movements without external references or heavy post-processing.

What would settle it

Capture a sequence of rapid self-occluding actions such as tumbling or object manipulation and verify whether the reconstructed pose stays drift-free and matches ground-truth markers over the full duration.

Figures

Figures reproduced from arXiv: 2604.07331 by Antonio Loquercio, Daniel Gehrig, Jefferson Ng, Luyang Hu, Wenjing Margaret Mao.

Figure 1
Figure 1. Figure 1: Illustration of RoSHI, a Robot-oriented Suit for Human Data In-the-Wild. RoSHI is a low-cost, portable system for in-the-wild human motion capture (bottom row), and deployment of learned policies on a humanoid robot (top row). On the left, the robot executes alternating single-leg jumps; on the right, it performs a bowing motion. RoSHI fuses signals from Project Aria glasses and nine Inertial Measurement U… view at source ↗
Figure 2
Figure 2. Figure 2: Overview of the RoSHI data pipeline. A user wears a low-cost, portable suit comprising nine IMU trackers (bottom [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Qualitative 3D articulated pose results of our method and various IMU-based and third-person view-based methods, [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Deployment of the learned humanoid policy on the [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
read the original abstract

Scaling up robot learning will likely require human data containing rich and long-horizon interactions in the wild. Existing approaches for collecting such data trade off portability, robustness to occlusion, and global consistency. We introduce RoSHI, a hybrid wearable that fuses low-cost sparse IMUs with the Project Aria glasses to estimate the full 3D pose and body shape of the wearer in a metric global coordinate frame from egocentric perception. This system is motivated by the complementarity of the two sensors: IMUs provide robustness to occlusions and high-speed motions, while egocentric SLAM anchors long-horizon motion and stabilizes upper body pose. We collect a dataset of agile activities to evaluate RoSHI. On this dataset, we generally outperform other egocentric baselines and perform comparably to a state-of-the-art exocentric baseline (SAM3D). Finally, we demonstrate that the motion data recorded from our system are suitable for real-world humanoid policy learning. For videos, data and more, visit the project webpage: https://roshi-mocap.github.io/

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The paper introduces RoSHI, a hybrid wearable system fusing low-cost sparse IMUs with Project Aria glasses for estimating full 3D human pose and body shape in a metric global coordinate frame from egocentric perception. It motivates the design via complementarity of IMUs (occlusion/high-speed robustness) and egocentric SLAM (long-horizon anchoring), collects a new dataset of agile activities, claims to generally outperform egocentric baselines while matching the exocentric SAM3D baseline, and demonstrates downstream utility for real-world humanoid policy learning.

Significance. If the quantitative results and policy-learning demonstration hold, this could be a meaningful contribution to scalable in-the-wild human motion capture for robot learning, offering a portable alternative that avoids heavy external infrastructure or post-processing. The hybrid sensor fusion and policy transfer experiment are practical strengths that directly address the paper's stated motivation.

major comments (2)
  1. [Abstract] Abstract: the claims of outperforming egocentric baselines and performing comparably to SAM3D are load-bearing for the central empirical contribution, yet the abstract (and available text) provides no quantitative metrics, error bars, dataset size, number of subjects/sequences, or statistical details; without these, the evaluation claims cannot be assessed.
  2. [Evaluation] Evaluation section (implied by dataset and baseline comparisons): the complementarity of IMUs and SLAM for agile real-world activities is asserted but requires explicit per-activity error breakdowns or ablation results to confirm robustness without external references or heavy post-processing, as this underpins the system's claimed advantages.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback. We address the major comments below and will revise the manuscript to strengthen the presentation of our quantitative results and analysis.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the claims of outperforming egocentric baselines and performing comparably to SAM3D are load-bearing for the central empirical contribution, yet the abstract (and available text) provides no quantitative metrics, error bars, dataset size, number of subjects/sequences, or statistical details; without these, the evaluation claims cannot be assessed.

    Authors: We agree that the abstract should include quantitative metrics to make the central claims assessable. In the revised version, we will add key results (e.g., mean per-joint position error for pose and shape, dataset statistics including number of subjects, sequences, and total duration) along with references to error bars and statistical details from the evaluation section. The full manuscript already reports these comparisons in detail, but we will ensure the abstract is self-contained. revision: yes

  2. Referee: [Evaluation] Evaluation section (implied by dataset and baseline comparisons): the complementarity of IMUs and SLAM for agile real-world activities is asserted but requires explicit per-activity error breakdowns or ablation results to confirm robustness without external references or heavy post-processing, as this underpins the system's claimed advantages.

    Authors: We acknowledge the value of more granular evidence. The manuscript reports overall results on agile activities and includes ablation studies on the hybrid fusion. To explicitly demonstrate complementarity and robustness, we will add per-activity error breakdowns (e.g., for high-speed motions and long-horizon sequences) and expanded ablation tables isolating IMU-only, SLAM-only, and combined performance. This will better support the design without external references or post-processing. revision: yes

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper introduces a hybrid hardware/software system (RoSHI) that fuses sparse IMUs with Project Aria egocentric SLAM for full-body metric pose and shape estimation. Its central claims rest on empirical evaluation: a newly collected in-the-wild dataset of agile activities, quantitative comparisons against external egocentric baselines and the exocentric SAM3D method, and a downstream policy-learning demonstration. No equations, derivations, fitted parameters, or self-referential predictions appear; performance is assessed against independent external references rather than by construction from the system's own outputs. The complementarity assumption is stated as motivation but is not required to be perfect for the modest reported gains to hold. This is a standard empirical systems paper whose evidence chain is externally anchored.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Abstract provides no explicit equations or parameters; the central claim rests on the unproven assumption that IMU-SLAM fusion yields metric global accuracy in the wild.

axioms (1)
  • domain assumption IMUs and egocentric SLAM are complementary for robust full-body pose estimation under occlusion and long horizons
    Explicitly stated as motivation in the abstract.

pith-pipeline@v0.9.0 · 5499 in / 1200 out tokens · 78333 ms · 2026-05-10T17:42:47.541226+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

42 extracted references · 22 canonical work pages · 1 internal anchor

  1. [1]

    Xsens mvn: Full 6dof human motion tracking using miniature inertial sensors,

    D. Roetenberg, H. Luinge, P. Slyckeet al., “Xsens mvn: Full 6dof human motion tracking using miniature inertial sensors,”Xsens Motion Technologies BV , Tech. Rep, vol. 1, no. 2009, pp. 1–7, 2009

    2009

  2. [2]

    Humans in 4d: Reconstructing and tracking humans with transform- ers,

    S. Goel, G. Pavlakos, J. Rajasegaran, A. Kanazawa, and J. Malik, “Humans in 4d: Reconstructing and tracking humans with transform- ers,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 14 783–14 794

    2023

  3. [3]

    Reconstructing hands in 3d with transformers,

    G. Pavlakos, D. Shan, I. Radosavovic, A. Kanazawa, D. Fouhey, and J. Malik, “Reconstructing hands in 3d with transformers,” 2023. [Online]. Available: https://arxiv.org/abs/2312.05251

    work page arXiv 2023

  4. [4]

    Simple pose: Rethinking and improving a bottom-up approach for multi-person pose estimation,

    J. Li, W. Su, and Z. Wang, “Simple pose: Rethinking and improving a bottom-up approach for multi-person pose estimation,” inProceedings of the AAAI conference on artificial intelligence, vol. 34, no. 07, 2020, pp. 11 354–11 361

    2020

  5. [5]

    Universal Manipulation Interface: In-The-Wild Robot Teaching Without In-The-Wild Robots

    C. Chi, Z. Xu, C. Pan, E. Cousineau, B. Burchfiel, S. Feng, R. Tedrake, and S. Song, “Universal manipulation interface: In- the-wild robot teaching without in-the-wild robots,” 2024. [Online]. Available: https://arxiv.org/abs/2402.10329

    work page internal anchor Pith review arXiv 2024

  6. [6]

    On bringing robots home

    N. M. M. Shafiullah, A. Rai, H. Etukuru, Y . Liu, I. Misra, S. Chin- tala, and L. Pinto, “On bringing robots home,”arXiv preprint arXiv:2311.16098, 2023

    work page arXiv 2023

  7. [7]

    Humanplus: Humanoid shadowing and imitation from humans,

    Z. Fu, Q. Zhao, Q. Wu, G. Wetzstein, and C. Finn, “Humanplus: Humanoid shadowing and imitation from humans,” 2024. [Online]. Available: https://arxiv.org/abs/2406.10454

    work page arXiv 2024

  8. [8]

    Deepmimic: example-guided deep reinforcement learning of physics- based character skills,

    X. B. Peng, P. Abbeel, S. Levine, and M. van de Panne, “Deepmimic: example-guided deep reinforcement learning of physics- based character skills,”ACM Transactions on Graphics, vol. 37, no. 4, p. 1–14, Jul. 2018. [Online]. Available: http://dx.doi.org/10. 1145/3197517.3201311

    work page arXiv 2018

  9. [9]

    Asap: Aligning simulation and real-world physics for learning agile humanoid whole-body skills,

    T. He, J. Gao, W. Xiao, Y . Zhang, Z. Wang, J. Wang, Z. Luo, G. He, N. Sobanbab, C. Pan, Z. Yi, G. Qu, K. Kitani, J. Hodgins, L. J. Fan, Y . Zhu, C. Liu, and G. Shi, “Asap: Aligning simulation and real-world physics for learning agile humanoid whole-body skills,”

  10. [10]

    Available: https://arxiv.org/abs/2502.01143

    [Online]. Available: https://arxiv.org/abs/2502.01143

    work page arXiv

  11. [11]

    Egomimic: Scaling imitation learning via egocentric video, 2024

    S. Kareer, D. Patel, R. Punamiya, P. Mathur, S. Cheng, C. Wang, J. Hoffman, and D. Xu, “Egomimic: Scaling imitation learning via egocentric video,” 2024. [Online]. Available: https://arxiv.org/abs/ 2410.24221

    work page arXiv 2024

  12. [12]

    Emma: Scaling mobile manipulation via egocentric human data.arXiv preprint arXiv:2509.04443, 2025

    L. Y . Zhu, P. Kuppili, R. Punamiya, P. Aphiwetsa, D. Patel, S. Kareer, S. Ha, and D. Xu, “Emma: Scaling mobile manipulation via egocentric human data,”arXiv preprint arXiv:2509.04443, 2025

    work page arXiv 2025

  13. [13]

    SMPL: A skinned multi-person linear model,

    M. Loper, N. Mahmood, J. Romero, G. Pons-Moll, and M. J. Black, “SMPL: A skinned multi-person linear model,”ACM Trans. Graphics (Proc. SIGGRAPH Asia), vol. 34, no. 6, pp. 248:1–248:16, Oct. 2015

    2015

  14. [14]

    Project aria: A new tool for egocentric multi-modal ai research.arXiv preprint arXiv:2308.13561, 2023

    J. Engel, K. Somasundaram, M. Goesele, A. Sun, A. Gamino, A. Turner, A. Talattof, A. Yuan, B. Souti, B. Meredith, C. Peng, C. Sweeney, C. Wilson, D. Barnes, D. DeTone, D. Caruso, D. Valleroy, D. Ginjupalli, D. Frost, E. Miller, E. Mueggler, E. Oleinik, F. Zhang, G. Somasundaram, G. Solaira, H. Lanaras, H. Howard-Jenkins, H. Tang, H. J. Kim, J. Rivera, J. ...

    work page arXiv 2023

  15. [15]

    Slimevr-tracker-esp: Slimevr tracker firmware for esp32/esp8266 and different imus,

    SlimeVR, “Slimevr-tracker-esp: Slimevr tracker firmware for esp32/esp8266 and different imus,”GitHub. [Online]. Available: https: //github.com/SlimeVR/SlimeVR-Tracker-ESP/releases/tag/v0.5.4, 2025, release v0.5.4 (Feb. 17, 2025). Accessed: Sep. 14, 2025

    2025

  16. [16]

    Intel® realsense™ sr300 coded light depth camera,

    A. Zabatani, V . Surazhsky, E. Sperling, S. B. Moshe, O. Menashe, D. H. Silver, Z. Karni, A. M. Bronstein, M. M. Bronstein, and R. Kimmel, “Intel® realsense™ sr300 coded light depth camera,” IEEE transactions on pattern analysis and machine intelligence, vol. 42, no. 10, pp. 2333–2345, 2019

    2019

  17. [17]

    Depth ac- curacy analysis of the zed 2i stereo camera in an indoor environment,

    A. Abdelsalam, M. Mansour, J. Porras, and A. Happonen, “Depth ac- curacy analysis of the zed 2i stereo camera in an indoor environment,” Robotics and Autonomous Systems, vol. 179, p. 104753, 2024

    2024

  18. [18]

    Tracking people by predicting 3d appearance, location and pose,

    J. Rajasegaran, G. Pavlakos, A. Kanazawa, and J. Malik, “Tracking people by predicting 3d appearance, location and pose,” inProceedings of the IEEE/CVF conference on computer vision and pattern recogni- tion, 2022, pp. 2740–2749

    2022

  19. [19]

    Pare: Part attention regressor for 3d human body estimation,

    M. Kocabas, C.-H. P. Huang, O. Hilliges, and M. J. Black, “Pare: Part attention regressor for 3d human body estimation,” 2021. [Online]. Available: https://arxiv.org/abs/2104.08527

    work page arXiv 2021

  20. [20]

    Monocular, one-stage, regression of multiple 3d people,

    Y . Sun, Q. Bao, W. Liu, Y . Fu, M. J. Black, and T. Mei, “Monocular, one-stage, regression of multiple 3d people,” 2021. [Online]. Available: https://arxiv.org/abs/2008.12272

    work page arXiv 2021

  21. [21]

    Learning to reconstruct 3d human pose and shape via model-fitting in the loop,

    N. Kolotouros, G. Pavlakos, M. J. Black, and K. Daniilidis, “Learning to reconstruct 3d human pose and shape via model-fitting in the loop,” 2019. [Online]. Available: https://arxiv.org/abs/1909.12828

    work page arXiv 2019

  22. [22]

    Deep two-stream video inference for human body pose and shape estimation,

    Z. Li, B. Xu, H. Huang, C. Lu, and Y . Guo, “Deep two-stream video inference for human body pose and shape estimation,” 2021. [Online]. Available: https://arxiv.org/abs/2110.11680

    work page arXiv 2021

  23. [23]

    Sam 3d body: Robust full-body human mesh recovery.arXiv preprint arXiv:2602.15989,

    X. Yang, D. Kukreja, D. Pinkus, A. Sagar, T. Fan, J. Park, S. Shin, J. Cao, J. Liu, N. Ugrinovic, M. Feiszli, J. Malik, P. Dollar, and K. Kitani, “Sam 3d body: Robust full-body human mesh recovery,” arXiv preprint arXiv:2602.15989, 2026

    work page arXiv 2026

  24. [24]

    Glamr: Global occlusion-aware human mesh recovery with dynamic cameras,

    Y . Yuan, U. Iqbal, P. Molchanov, K. Kitani, and J. Kautz, “Glamr: Global occlusion-aware human mesh recovery with dynamic cameras,” 2022. [Online]. Available: https://arxiv.org/abs/2112.01524

    work page arXiv 2022

  25. [25]

    Ego4d: Around the world in 3,000 hours of egocentric video,

    K. Graumanet al., “Ego4d: Around the world in 3,000 hours of egocentric video,” 2022. [Online]. Available: https://arxiv.org/abs/ 2110.07058

    work page arXiv 2022

  26. [26]

    Scaling egocentric vision: The epic-kitchens dataset,

    Damenet al., “Scaling egocentric vision: The epic-kitchens dataset,” inEuropean Conference on Computer Vision (ECCV), 2018

    2018

  27. [27]

    The epic-kitchens dataset: Collection, challenges and baselines,

    D. Damen, H. Doughty, G. M. Farinella, S. Fidler, A. Furnari, E. Kazakos, D. Moltisanti, J. Munro, T. Perrett, W. Price, and M. Wray, “The epic-kitchens dataset: Collection, challenges and baselines,” 2020. [Online]. Available: https://arxiv.org/abs/2005.00343

    work page arXiv 2020

  28. [28]

    Karen Liu, Ziwei Liu, Jakob Engel, Renzo De Nardi, and Richard Newcombe

    L. Ma, Y . Ye, F. Hong, V . Guzov, Y . Jiang, R. Postyeni, L. Pesqueira, A. Gamino, V . Baiyya, H. J. Kim, K. Bailey, D. S. Fosas, C. K. Liu, Z. Liu, J. Engel, R. D. Nardi, and R. Newcombe, “Nymeria: A massive collection of multimodal egocentric daily motion in the wild,” 2024. [Online]. Available: https://arxiv.org/abs/2406.09905

    work page arXiv 2024

  29. [29]

    EgoLife: Towards egocentric life assistant.arXiv preprint arXiv:2503.03803, 2025

    J. Yanget al., “Egolife: Towards egocentric life assistant,” 2025. [Online]. Available: https://arxiv.org/abs/2503.03803

    work page arXiv 2025

  30. [30]

    Estimating body and hand motion in an ego- sensed world,

    B. Yi, V . Ye, M. Zheng, Y . Li, L. M¨uller, G. Pavlakos, Y . Ma, J. Malik, and A. Kanazawa, “Estimating body and hand motion in an ego- sensed world,” inProceedings of the Computer Vision and Pattern Recognition Conference, 2025, pp. 7072–7084

    2025

  31. [31]

    Optitrack motion capture systems,

    OptiTrack, “Optitrack motion capture systems,” https://optitrack.com/, accessed: 2026-03-01

    2026

  32. [32]

    Vicon: Motion capture systems,

    Vicon, “Vicon: Motion capture systems,” https://www.vicon.com/, accessed: 2026-03-01

    2026

  33. [33]

    [Online]

    Noitom International Limited, “Noitom,”Website. [Online]. Available: https://www.noitom.com/, 2026, accessed: Feb. 23, 2026

    2026

  34. [34]

    Accurate human motion capture in large areas by combining imu- and laser-based people tracking,

    J. Ziegler, H. Kretzschmar, C. Stachniss, G. Grisetti, and W. Burgard, “Accurate human motion capture in large areas by combining imu- and laser-based people tracking,”2011 IEEE/RSJ International Conference on Intelligent Robots and Systems, pp. 86–91, 2011. [Online]. Available: https://api.semanticscholar.org/CorpusID:1505190

    2011

  35. [35]

    Amass: Archive of motion capture as surface shapes,

    N. Mahmood, N. Ghorbani, N. F. Troje, G. Pons-Moll, and M. J. Black, “Amass: Archive of motion capture as surface shapes,” inProceedings of the IEEE/CVF international conference on computer vision, 2019, pp. 5442–5451

    2019

  36. [36]

    Gfpose: Learning 3d human pose prior with gradient fields,

    H. Ci, M. Wu, W. Zhu, X. Ma, H. Dong, F. Zhong, and Y . Wang, “Gfpose: Learning 3d human pose prior with gradient fields,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2023, pp. 4800–4810

    2023

  37. [37]

    Probabilistic human mesh recovery in 3d scenes from egocentric views,

    S. Zhang, Q. Ma, Y . Zhang, S. Aliakbarian, D. Cosker, and S. Tang, “Probabilistic human mesh recovery in 3d scenes from egocentric views,” inProceedings of the IEEE/CVF International Conference on Computer Vision, 2023, pp. 7989–8000

    2023

  38. [38]

    AprilTag: A robust and flexible visual fiducial system,

    E. Olson, “AprilTag: A robust and flexible visual fiducial system,” in Proceedings of the IEEE International Conference on Robotics and Automation (ICRA). IEEE, May 2011, pp. 3400–3407

    2011

  39. [39]

    Expressive body capture: 3d hands, face, and body from a single image,

    G. Pavlakos, V . Choutas, N. Ghorbani, T. Bolkart, A. A. A. Osman, D. Tzionas, and M. J. Black, “Expressive body capture: 3d hands, face, and body from a single image,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2019

    2019

  40. [40]

    Jihoon Kim, Taehyun Byun, Seungyoun Shin, Jungdam Won, and Sungjoon Choi

    C. M. Kim*, B. Yi*, H. Choi, Y . Ma, K. Goldberg, and A. Kanazawa, “Pyroki: A modular toolkit for robot kinematic optimization,” 2025. [Online]. Available: https://arxiv.org/abs/2505.03728

    work page arXiv 2025

  41. [41]

    Retargeting Matters: General Motion Retarget- ing for Humanoid Motion Tracking,

    J. P. Araujo, Y . Ze, P. Xu, J. Wu, and C. K. Liu, “Retargeting matters: General motion retargeting for humanoid motion tracking,” arXiv preprint arXiv:2510.02252, 2025

    work page arXiv 2025

  42. [42]

    Beyondmimic: From mo- tion tracking to versatile humanoid control via guided diffusion,

    Q. Liao, T. E. Truong, X. Huang, G. Tevet, K. Sreenath, and C. K. Liu, “Beyondmimic: From motion tracking to versatile humanoid control via guided diffusion,” 2025. [Online]. Available: https://arxiv.org/abs/2508.08241

    work page arXiv 2025