Markerless Motion Capture for Biomechanical Whole-Body Kinematic Estimation in Infants

Colleen Peyton; Divya Joshi; J.D. Peiffer; R. James Cotton

arxiv: 2605.17120 · v1 · pith:UQ2X753Knew · submitted 2026-05-16 · 💻 cs.CV

Markerless Motion Capture for Biomechanical Whole-Body Kinematic Estimation in Infants

Divya Joshi , J.D. Peiffer , Colleen Peyton , R. James Cotton This is my paper

Pith reviewed 2026-05-20 15:21 UTC · model grok-4.3

classification 💻 cs.CV

keywords infant biomechanicsmarkerless motion capture3D pose estimationmotor developmentinverse kinematicsvideo-based assessmentspontaneous movement

0 comments

The pith

Biomechanical models fitted to 3D pose estimates from video distinguish infant movement patterns related to motor development.

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

The paper evaluates three state-of-the-art 3D pose estimation methods on 100 videos from eight infants captured with a multi-view markerless system. It quantifies accuracy through reprojection error, geometric consistency, and 3D position error, then fits an inverse kinematic model to the outputs of the strongest method for whole-body kinematic reconstruction. In a case comparison, the resulting models separate representative movement patterns that a clinical expert associates with motor development status. This tests whether video alone can support objective, scalable checks for early motor impairment that now rely on direct expert observation of spontaneous movement.

Core claim

The central claim is that SAM 3D Body supplies the most useful 3D joint positions among the tested frameworks for fitting inverse kinematics to infant data, with Procrustes-aligned position errors of 19 to 28 mm, and that biomechanical models built from those estimates distinguish representative movement patterns tied to motor development in a clinical case comparison.

What carries the argument

SAM 3D Body 3D pose estimation, which produces joint positions that are then fed into an inverse kinematic framework to reconstruct full-body biomechanical motion.

If this is right

Video from ordinary cameras can support objective quantification of spontaneous infant movement without attached markers.
Trade-offs among current pose estimators become measurable for infant-specific data through reprojection and position metrics.
Proof-of-concept kinematic fitting opens the door to automated screening that supplements expert visual assessment.

Where Pith is reading between the lines

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

The method could be validated on larger cohorts to test whether early kinematic differences predict later motor milestones.
Home-recorded videos might allow remote monitoring if accuracy holds under less controlled conditions.
Hybrid use of multiple pose estimators could reduce errors for the small body proportions typical in infancy.

Load-bearing premise

The multi-view markerless motion capture system produces 3D positions accurate enough to serve as ground truth for measuring the pose estimators and for supporting inverse kinematic fitting on infant data.

What would settle it

A larger study in which independent clinical ratings or longitudinal motor outcomes show no reliable match with the movement distinctions produced by the biomechanical models fitted to the video pose estimates.

Figures

Figures reproduced from arXiv: 2605.17120 by Colleen Peyton, Divya Joshi, J.D. Peiffer, R. James Cotton.

**Figure 1.** Figure 1: The MMMC experimental setup. The infant is placed supine on a mat in the center of 8 tripods with cameras pointing at the mat. 2.3 Evaluation Dataset The dataset evaluated in this study comprised the following: • Participants: 8 infants, 5 to 16 weeks post-menstrual age • Sessions: 13 unique sessions across all participants • Cameras: 8 synchronized, calibrated views per session • Recordings: 124 multi-vie… view at source ↗

**Figure 2.** Figure 2: Example of detected and reprojected 2D keypoints across methods. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** Figure 3: Per keypoint reprojection error and position error across keypoint detection [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

**Figure 4.** Figure 4: Example of monocular kinematic reconstruction from SAM 3D Body (MoVi 87) [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗

read the original abstract

arly identification of motor impairment in infancy relies on expert visual assessment of spontaneous movement, motivating the development of automated, objective alternatives. One promising approach is using computer vision, which benefits from high quality pose estimation from video. In this study, we systematically evaluated three state-of-the-art pose estimation frameworks (MeTRAbs-ACAE, SAM 3D Body, and Sapiens) on 100 videos over 13 sessions of 8 infants recorded with a multi-view markerless motion capture system. We quantified keypoint detection accuracy using reprojection error, geometric consistency, and Procrustes-aligned 3D position error, and demonstrated proof-of-concept for fitting an inverse kinematic framework to infant data. While Sapiens achieved the lowest reprojection error and highest geometric consistency of the methods evaluated (22.8 pixels and 0.82, respectively), SAM 3D Body provided the most comprehensive 3D information for kinematic reconstruction with Procrustes-aligned position errors of 19 to 28 mm. We demonstrate in a case comparison example that biomechanical models fit to SAM 3D estimates distinguish representative movement patterns in infants related to motor development, as identified by a clinical expert. Together, these findings highlight both the promise and current limitations of 3D pose estimation for infant biomechanics and establish preliminary groundwork for scalable, video-based assessment of early motor development.

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.

Desk Editor's Note private letter to a colleague

The paper gives solid error benchmarks for pose estimation on real infant videos but the main kinematic claim is a single unquantified case example.

read the letter

This paper evaluates three pose estimation methods on infant videos and takes a first step toward using the outputs for whole-body inverse kinematics. The concrete contribution is the error numbers from a multi-view capture setup on real infant data. They tested MeTRAbs-ACAE, SAM 3D Body, and Sapiens across 100 videos from 8 infants. Sapiens had the best reprojection and consistency scores, while SAM 3D gave the most usable 3D keypoints with 19-28 mm Procrustes error. They then fit a biomechanical model and showed in a single case that the kinematics could separate movement patterns noted by a clinical expert. The new part is applying this pipeline specifically to infants and closing the loop to biomechanical estimates rather than just reporting keypoints. The multi-view system provides a useful external validation that many video-only studies lack. The evaluation is honest about the limitations and gives numbers that others can build on. It is the kind of applied work that fills a gap in the literature for early motor development. The soft spot is that the distinguishing claim rests on one qualitative example without quantitative kinematic metrics or controls. There is no report of how the joint angles differed, no statistical separation, and no test of whether the same distinction appears when fitting to the mocap positions or to perturbed data. With millimeter-scale errors on short infant limbs, angle errors could be substantial, but that is not checked. The methods section also leaves out details on data selection and the exact IK setup. This is relevant for researchers in pediatric biomechanics and computer vision for health. The evaluation section is solid enough that it should go to peer review, though the authors will likely need to add quantitative validation for the kinematics part to make the clinical angle convincing.

Referee Report

1 major / 1 minor

Summary. The paper evaluates three state-of-the-art 3D pose estimation frameworks (MeTRAbs-ACAE, SAM 3D Body, and Sapiens) on 100 videos from 13 sessions of 8 infants recorded with a multi-view markerless motion capture system. Accuracy is quantified via reprojection error, geometric consistency, and Procrustes-aligned 3D position error, with SAM 3D Body identified as providing the most useful 3D information (19–28 mm errors). The manuscript presents a proof-of-concept for fitting an inverse kinematic biomechanical model to the SAM 3D estimates and demonstrates in a single case comparison example that the resulting models distinguish representative infant movement patterns related to motor development as identified by a clinical expert.

Significance. If the core results hold under more rigorous validation, the work could support development of scalable, video-based tools for objective assessment of early motor development in infants, addressing a clinically important need for early detection of impairments. The systematic benchmarking of multiple pose estimators on real infant data and the linkage to whole-body biomechanical modeling are strengths that could inform future research in computer vision applications to pediatrics.

major comments (1)

[Case comparison example (results section)] The central claim that biomechanical models fit to SAM 3D estimates distinguish representative movement patterns relies on a single qualitative case comparison example without quantitative kinematic metrics (e.g., joint angle ranges, velocities, smoothness, or coordination scores), statistical separation, or control comparisons to the multi-view mocap 3D positions or to perturbed keypoints. Given the 19–28 mm Procrustes-aligned errors reported for infant-scale limbs, an analysis of error propagation into derived joint angles or other kinematic quantities is needed to establish that the observed distinctions are robust rather than artifacts.

minor comments (1)

[Abstract] The abstract would benefit from brief additional details on data exclusion criteria, the specific inverse kinematic fitting procedures used, and any error propagation steps performed.

Simulated Author's Rebuttal

1 responses · 1 unresolved

We thank the referee for their constructive and detailed feedback on our manuscript. The comments highlight important considerations for strengthening the proof-of-concept demonstration, and we have revised the manuscript to address them where possible while preserving the scope of this preliminary study.

read point-by-point responses

Referee: The central claim that biomechanical models fit to SAM 3D estimates distinguish representative movement patterns relies on a single qualitative case comparison example without quantitative kinematic metrics (e.g., joint angle ranges, velocities, smoothness, or coordination scores), statistical separation, or control comparisons to the multi-view mocap 3D positions or to perturbed keypoints. Given the 19–28 mm Procrustes-aligned errors reported for infant-scale limbs, an analysis of error propagation into derived joint angles or other kinematic quantities is needed to establish that the observed distinctions are robust rather than artifacts.

Authors: We agree that the demonstration is limited to a single qualitative case comparison and lacks quantitative kinematic metrics, statistical tests, and explicit error propagation analysis. This was presented as an initial proof-of-concept to illustrate feasibility given the practical constraints of collecting multi-view markerless data from infants. In the revised manuscript, we have added quantitative kinematic metrics (joint angle ranges and movement smoothness indices) for the two cases and included direct comparisons of the derived biomechanical quantities against the multi-view mocap ground-truth positions. We have also incorporated a simple error propagation estimate showing that the reported 19–28 mm 3D position errors correspond to approximately 5–9° uncertainty in joint angles for typical infant limb segment lengths, which is consistent with the gross pattern distinctions noted by the clinical expert. We acknowledge that a full sensitivity analysis using perturbed keypoints and larger-scale statistical separation would further strengthen the claims but is beyond the current dataset size and is noted as a direction for future work. revision: partial

standing simulated objections not resolved

Full statistical separation across multiple subjects with formal hypothesis testing, as this requires a substantially larger cohort of infant recordings with corresponding clinical assessments that are not available in the present study.

Circularity Check

0 steps flagged

No significant circularity detected

full rationale

The paper's evaluation chain relies on external, independent metrics (reprojection error, geometric consistency, Procrustes-aligned 3D position error) computed against multi-view markerless mocap ground truth, plus a qualitative case comparison distinguished by a clinical expert label. No equations or steps reduce a claimed prediction or kinematic output to a fitted parameter by construction, and no self-citations are invoked as load-bearing uniqueness theorems or ansatzes for the core claims. The derivation remains self-contained against external benchmarks and does not rename known results or smuggle assumptions via prior author work.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The evaluation depends on the domain assumption that adult-trained pose models generalize to infants and that the multi-view system serves as reliable ground truth for error quantification.

axioms (1)

domain assumption Pose estimation models trained primarily on adult data can be applied to infant videos with meaningful accuracy for biomechanical purposes.
Infant proportions and spontaneous movements differ substantially from adults, yet the study proceeds with direct application.

pith-pipeline@v0.9.0 · 5782 in / 1215 out tokens · 56972 ms · 2026-05-20T15:21:05.076690+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

36 extracted references · 36 canonical work pages

[1]

2017 , journal =

Novak, Iona and others , number =. 2017 , journal =. doi:10.1001/jamapediatrics.2017.1689 , issn =

work page doi:10.1001/jamapediatrics.2017.1689 2017
[2]

2015 , journal =

Herskind, Anna and Greisen, Gorm and Nielsen, Jens Bo , number =. 2015 , journal =. doi:10.1111/dmcn.12531 , issn =

work page doi:10.1111/dmcn.12531 2015
[3]

2013 , journal =

Bosanquet, Margot and Copeland, Lisa and Ware, Robert and Boyd, Roslyn , number =. 2013 , journal =. doi:10.1111/dmcn.12140 , issn =

work page doi:10.1111/dmcn.12140 2013
[4]

and Doyle, Lex W

Spittle, Alicia J. and Doyle, Lex W. and Boyd, Roslyn N. , number =. 2008 , journal =. doi:10.1111/j.1469-8749.2008.02025.x , issn =

work page doi:10.1111/j.1469-8749.2008.02025.x 2008
[5]

2024 , journal =

Peyton, Colleen and others , publisher =. 2024 , journal =. doi:10.1111/dmcn.15925 , issn =

work page doi:10.1111/dmcn.15925 2024
[6]

2024 , journal =

Barbosa, Vanessa Maziero and Peyton, Colleen and Sukal‐Moulton, Theresa , publisher =. 2024 , journal =. doi:10.1111/dmcn.15926 , issn =

work page doi:10.1111/dmcn.15926 2024
[7]

2012 , journal =

Noble, Yolande and Boyd, Roslyn , number =. 2012 , journal =. doi:10.1111/j.1469-8749.2010.03903.x , issn =

work page doi:10.1111/j.1469-8749.2010.03903.x 2012
[8]

and Arens, Michael and Hofmann, Ulrich G

Hesse, Nikolas and Pujades, Sergi and Black, Michael J. and Arens, Michael and Hofmann, Ulrich G. and Schroeder, A. Sebastian , number =. 2020 , journal =. doi:10.1109/TPAMI.2019.2917908 , issn =

work page doi:10.1109/tpami.2019.2917908 2020
[9]

2020 , journal =

Chambers, Claire and others , number =. 2020 , journal =. doi:10.1109/TNSRE.2020.3029121 , issn =

work page doi:10.1109/tnsre.2020.3029121 2020
[10]

and Johnson, Michelle J

Segado, Melanie and Prosser, Laura and Duncan, Andrea F. and Johnson, Michelle J. and Kording, Konrad P. , month =. 2025 , journal =

work page 2025
[11]

and Shah, Kunal and Miller, Lee E and Cotton, R

Peiffer, J.D. and Shah, Kunal and Anarwala, Shawana and Abdou, Kayan and Cotton, R. James , urldate =. Fusing Uncalibrated. 2024 10th. doi:10.1109/BioRob60516.2024.10719724 , eventtitle =

work page doi:10.1109/biorob60516.2024.10719724 2024
[12]

James , urldate =

Cotton, R. James , urldate =. Differentiable Biomechanics Unlocks Opportunities for Markerless Motion Capture , issn =. 2025 International Conference On Rehabilitation Robotics (. doi:10.1109/ICORR66766.2025.11063174 , eventtitle =

work page doi:10.1109/icorr66766.2025.11063174 2025
[13]

2025 , arxivId =

Gama, Filipe and others , journal =. 2025 , arxivId =

work page 2025
[14]

, number =

Ali, Manal Mostafa and Mohamed, Samer I. , number =. 2024 , journal =. doi:10.1007/S11042-024-19198-5/FIGURES/16 , issn =

work page doi:10.1007/s11042-024-19198-5/figures/16 2024
[15]

2025 , journal =

Jahn, Lennart and others , number =. 2025 , journal =. doi:10.1038/s41598-025-96206-0 , issn =

work page doi:10.1038/s41598-025-96206-0 2025
[16]

2022 , journal =

Groos, Daniel and others , number =. 2022 , journal =. doi:10.1001/jamanetworkopen.2022.21325 , issn =

work page doi:10.1001/jamanetworkopen.2022.21325 2022
[17]

2014 , journal =

Cahill-Rowley, Katelyn and Rose, Jessica , number =. 2014 , journal =. doi:10.1111/DMCN.12355 , issn =

work page doi:10.1111/dmcn.12355 2014
[18]

Learning 3D Human Pose Estimation From Dozens of Datasets Using a Geometry-Aware Autoencoder To Bridge Between Skeleton Formats , booktitle =

Sárándi, István and Hermans, Alexander and Leibe, Bastian , urldate =. Learning 3D Human Pose Estimation From Dozens of Datasets Using a Geometry-Aware Autoencoder To Bridge Between Skeleton Formats , booktitle =

work page
[19]

2023 , journal =

Mohammed, Shawan and Siebers, Hannah and Preu. 2023 , journal =

work page 2023
[20]

ArXiv , author =

Markerless Motion Capture and Biomechanical Analysis Pipeline , year =. ArXiv , author =

work page
[21]

and Shah, Kunal and Miller, Lee E and Cotton, R

Firouzabadi, Pouyan and Murray, Wendy and Sobinov, Anton R and Peiffer, J.D. and Shah, Kunal and Miller, Lee E and Cotton, R. James , urldate =. Biomechanical Arm and Hand Tracking with Multiview Markerless Motion Capture , issn =. 2024 10th. doi:10.1109/BioRob60516.2024.10719940 , eventtitle =

work page doi:10.1109/biorob60516.2024.10719940 2024
[22]

doi:10.1016/j.apmr.2022.08.868 , shorttitle =

Archives of Physical Medicine and Rehabilitation , author =. doi:10.1016/j.apmr.2022.08.868 , shorttitle =

work page doi:10.1016/j.apmr.2022.08.868 2022
[23]

2021 , journal =

Karashchuk, Pierre and others , number =. 2021 , journal =. doi:10.1016/j.celrep.2021.109730 , issn =

work page doi:10.1016/j.celrep.2021.109730 2021
[24]

, author =

EasyMoCap - Make human motion capture easier. , author =. 2021 , url =

work page 2021
[25]

2021 , journal =

Ghorbani, Saeed and others , number =. 2021 , journal =. doi:10.1371/journal.pone.0253157 , issn =

work page doi:10.1371/journal.pone.0253157 2021
[26]

2025 , journal =

Ferguson, Aaron and others , month =. 2025 , journal =

work page 2025
[27]

2022 , booktitle =

Roy, Soumava Kumar and Citraro, Leonardo and Honari, Sina and Fua, Pascal , doi =. 2022 , booktitle =

work page 2022
[28]

James and Cimorelli, Anthony and Shah, Kunal and Anarwala, Shawana and Uhlrich, Scott and Karakostas, Tasos , urldate =

Cotton, R. James and Cimorelli, Anthony and Shah, Kunal and Anarwala, Shawana and Uhlrich, Scott and Karakostas, Tasos , urldate =. Improved Trajectory Reconstruction for Markerless Pose Estimation , issn =. 2023 45th Annual International Conference of the. doi:10.1109/EMBC40787.2023.10340745 , eventtitle =

work page doi:10.1109/embc40787.2023.10340745 2023
[29]

2025 , journal =

Unger, Tim and others , doi =. 2025 , journal =

work page 2025
[30]

Mental Retardation and Developmental Disabilities Research Reviews , author=

Prechtl's assessment of general movements: A diagnostic tool for the functional assessment of the young nervous system , year=. Mental Retardation and Developmental Disabilities Research Reviews , author=

work page
[31]

2024 , journal =

Sukal‐Moulton, Theresa and others , publisher =. 2024 , journal =. doi:10.1111/dmcn.15924 , issn =

work page doi:10.1111/dmcn.15924 2024
[32]

2012 , booktitle =

Todorov, Emanuel and Erez, Tom and Tassa, Yuval , doi =. 2012 , booktitle =

work page 2012
[33]

James and others , doi =

Cotton, R. James and others , doi =. 2023 , booktitle =

work page 2023
[34]

2026 , journal =

Cotton, R James and Firouzabadi, Pouyan and Murray, Wendy , arxivId =. 2026 , journal =

work page 2026
[35]

2026 , issn =

Effect of markers in training dataset for markerless applications in biomechanics , journal =. 2026 , issn =. doi:https://doi.org/10.1016/j.jbiomech.2026.113341 , author =

work page doi:10.1016/j.jbiomech.2026.113341 2026
[36]

2006 , doi =

Pediatric interventional radiography equipment: safety considerations , journal =. 2006 , doi =

work page 2006

[1] [1]

2017 , journal =

Novak, Iona and others , number =. 2017 , journal =. doi:10.1001/jamapediatrics.2017.1689 , issn =

work page doi:10.1001/jamapediatrics.2017.1689 2017

[2] [2]

2015 , journal =

Herskind, Anna and Greisen, Gorm and Nielsen, Jens Bo , number =. 2015 , journal =. doi:10.1111/dmcn.12531 , issn =

work page doi:10.1111/dmcn.12531 2015

[3] [3]

2013 , journal =

Bosanquet, Margot and Copeland, Lisa and Ware, Robert and Boyd, Roslyn , number =. 2013 , journal =. doi:10.1111/dmcn.12140 , issn =

work page doi:10.1111/dmcn.12140 2013

[4] [4]

and Doyle, Lex W

Spittle, Alicia J. and Doyle, Lex W. and Boyd, Roslyn N. , number =. 2008 , journal =. doi:10.1111/j.1469-8749.2008.02025.x , issn =

work page doi:10.1111/j.1469-8749.2008.02025.x 2008

[5] [5]

2024 , journal =

Peyton, Colleen and others , publisher =. 2024 , journal =. doi:10.1111/dmcn.15925 , issn =

work page doi:10.1111/dmcn.15925 2024

[6] [6]

2024 , journal =

Barbosa, Vanessa Maziero and Peyton, Colleen and Sukal‐Moulton, Theresa , publisher =. 2024 , journal =. doi:10.1111/dmcn.15926 , issn =

work page doi:10.1111/dmcn.15926 2024

[7] [7]

2012 , journal =

Noble, Yolande and Boyd, Roslyn , number =. 2012 , journal =. doi:10.1111/j.1469-8749.2010.03903.x , issn =

work page doi:10.1111/j.1469-8749.2010.03903.x 2012

[8] [8]

and Arens, Michael and Hofmann, Ulrich G

Hesse, Nikolas and Pujades, Sergi and Black, Michael J. and Arens, Michael and Hofmann, Ulrich G. and Schroeder, A. Sebastian , number =. 2020 , journal =. doi:10.1109/TPAMI.2019.2917908 , issn =

work page doi:10.1109/tpami.2019.2917908 2020

[9] [9]

2020 , journal =

Chambers, Claire and others , number =. 2020 , journal =. doi:10.1109/TNSRE.2020.3029121 , issn =

work page doi:10.1109/tnsre.2020.3029121 2020

[10] [10]

and Johnson, Michelle J

Segado, Melanie and Prosser, Laura and Duncan, Andrea F. and Johnson, Michelle J. and Kording, Konrad P. , month =. 2025 , journal =

work page 2025

[11] [11]

and Shah, Kunal and Miller, Lee E and Cotton, R

Peiffer, J.D. and Shah, Kunal and Anarwala, Shawana and Abdou, Kayan and Cotton, R. James , urldate =. Fusing Uncalibrated. 2024 10th. doi:10.1109/BioRob60516.2024.10719724 , eventtitle =

work page doi:10.1109/biorob60516.2024.10719724 2024

[12] [12]

James , urldate =

Cotton, R. James , urldate =. Differentiable Biomechanics Unlocks Opportunities for Markerless Motion Capture , issn =. 2025 International Conference On Rehabilitation Robotics (. doi:10.1109/ICORR66766.2025.11063174 , eventtitle =

work page doi:10.1109/icorr66766.2025.11063174 2025

[13] [13]

2025 , arxivId =

Gama, Filipe and others , journal =. 2025 , arxivId =

work page 2025

[14] [14]

, number =

Ali, Manal Mostafa and Mohamed, Samer I. , number =. 2024 , journal =. doi:10.1007/S11042-024-19198-5/FIGURES/16 , issn =

work page doi:10.1007/s11042-024-19198-5/figures/16 2024

[15] [15]

2025 , journal =

Jahn, Lennart and others , number =. 2025 , journal =. doi:10.1038/s41598-025-96206-0 , issn =

work page doi:10.1038/s41598-025-96206-0 2025

[16] [16]

2022 , journal =

Groos, Daniel and others , number =. 2022 , journal =. doi:10.1001/jamanetworkopen.2022.21325 , issn =

work page doi:10.1001/jamanetworkopen.2022.21325 2022

[17] [17]

2014 , journal =

Cahill-Rowley, Katelyn and Rose, Jessica , number =. 2014 , journal =. doi:10.1111/DMCN.12355 , issn =

work page doi:10.1111/dmcn.12355 2014

[18] [18]

Learning 3D Human Pose Estimation From Dozens of Datasets Using a Geometry-Aware Autoencoder To Bridge Between Skeleton Formats , booktitle =

Sárándi, István and Hermans, Alexander and Leibe, Bastian , urldate =. Learning 3D Human Pose Estimation From Dozens of Datasets Using a Geometry-Aware Autoencoder To Bridge Between Skeleton Formats , booktitle =

work page

[19] [19]

2023 , journal =

Mohammed, Shawan and Siebers, Hannah and Preu. 2023 , journal =

work page 2023

[20] [20]

ArXiv , author =

Markerless Motion Capture and Biomechanical Analysis Pipeline , year =. ArXiv , author =

work page

[21] [21]

and Shah, Kunal and Miller, Lee E and Cotton, R

Firouzabadi, Pouyan and Murray, Wendy and Sobinov, Anton R and Peiffer, J.D. and Shah, Kunal and Miller, Lee E and Cotton, R. James , urldate =. Biomechanical Arm and Hand Tracking with Multiview Markerless Motion Capture , issn =. 2024 10th. doi:10.1109/BioRob60516.2024.10719940 , eventtitle =

work page doi:10.1109/biorob60516.2024.10719940 2024

[22] [22]

doi:10.1016/j.apmr.2022.08.868 , shorttitle =

Archives of Physical Medicine and Rehabilitation , author =. doi:10.1016/j.apmr.2022.08.868 , shorttitle =

work page doi:10.1016/j.apmr.2022.08.868 2022

[23] [23]

2021 , journal =

Karashchuk, Pierre and others , number =. 2021 , journal =. doi:10.1016/j.celrep.2021.109730 , issn =

work page doi:10.1016/j.celrep.2021.109730 2021

[24] [24]

, author =

EasyMoCap - Make human motion capture easier. , author =. 2021 , url =

work page 2021

[25] [25]

2021 , journal =

Ghorbani, Saeed and others , number =. 2021 , journal =. doi:10.1371/journal.pone.0253157 , issn =

work page doi:10.1371/journal.pone.0253157 2021

[26] [26]

2025 , journal =

Ferguson, Aaron and others , month =. 2025 , journal =

work page 2025

[27] [27]

2022 , booktitle =

Roy, Soumava Kumar and Citraro, Leonardo and Honari, Sina and Fua, Pascal , doi =. 2022 , booktitle =

work page 2022

[28] [28]

James and Cimorelli, Anthony and Shah, Kunal and Anarwala, Shawana and Uhlrich, Scott and Karakostas, Tasos , urldate =

Cotton, R. James and Cimorelli, Anthony and Shah, Kunal and Anarwala, Shawana and Uhlrich, Scott and Karakostas, Tasos , urldate =. Improved Trajectory Reconstruction for Markerless Pose Estimation , issn =. 2023 45th Annual International Conference of the. doi:10.1109/EMBC40787.2023.10340745 , eventtitle =

work page doi:10.1109/embc40787.2023.10340745 2023

[29] [29]

2025 , journal =

Unger, Tim and others , doi =. 2025 , journal =

work page 2025

[30] [30]

Mental Retardation and Developmental Disabilities Research Reviews , author=

Prechtl's assessment of general movements: A diagnostic tool for the functional assessment of the young nervous system , year=. Mental Retardation and Developmental Disabilities Research Reviews , author=

work page

[31] [31]

2024 , journal =

Sukal‐Moulton, Theresa and others , publisher =. 2024 , journal =. doi:10.1111/dmcn.15924 , issn =

work page doi:10.1111/dmcn.15924 2024

[32] [32]

2012 , booktitle =

Todorov, Emanuel and Erez, Tom and Tassa, Yuval , doi =. 2012 , booktitle =

work page 2012

[33] [33]

James and others , doi =

Cotton, R. James and others , doi =. 2023 , booktitle =

work page 2023

[34] [34]

2026 , journal =

Cotton, R James and Firouzabadi, Pouyan and Murray, Wendy , arxivId =. 2026 , journal =

work page 2026

[35] [35]

2026 , issn =

Effect of markers in training dataset for markerless applications in biomechanics , journal =. 2026 , issn =. doi:https://doi.org/10.1016/j.jbiomech.2026.113341 , author =

work page doi:10.1016/j.jbiomech.2026.113341 2026

[36] [36]

2006 , doi =

Pediatric interventional radiography equipment: safety considerations , journal =. 2006 , doi =

work page 2006