Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

ATLAS is a parametric 3D human body model that keeps the internal skeleton and external soft tissue as separate, independent control axes.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

ATLAS decouples skeleton and shape parameters in a parametric human body model, improving fit accuracy and controllability over previous models like SMPL-X.

T0 review reviewed 2026-08-05 challenge →

load-bearing objection Strong industrial body model with a real architectural decoupling, but the keypoint-based skeleton fitting makes the decoupling claim conditional until proven on anatomical data. the 3 major comments →

arxiv 2508.15767 v1 pith:UXRVQ6LW submitted 2025-08-21 cs.CV

ATLAS: Decoupling Skeletal and Shape Parameters for Expressive Parametric Human Modeling

Jinhyung Park , Javier Romero , Shunsuke Saito , Fabian Prada , Takaaki Shiratori , Yichen Xu , Federica Bogo , Shoou-I Yu
show 2 more authors
Kris Kitani Rawal Khirodkar
This is my paper
classification cs.CV
keywords parametric human body modelskeleton-shape decouplinglinear blend skinningpose-dependent deformations3D human mesh fittingsingle-image human reconstructionskeletal attributes
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The reading

ATLAS is a 3D human body model that treats the internal skeleton and the outer soft tissue as two separate parameter sets, instead of deriving joint positions from surface vertices as earlier models do. The surface is first reshaped with a linear basis while attached to a fixed template skeleton; then 76 skeletal attributes—15 body-part scales and 61 bone lengths—scale and pose the whole mesh through linear blend skinning. Because the skeleton never changes when the surface changes, an artist can widen shoulders or lengthen arms without disturbing body weight, and a fitting algorithm can use keypoints to fix the skeleton without hallucinating fat or muscle. Trained on 600k high-resolution scans, ATLAS reports lower mean vertex error than the leading vertex-centric model on held-out scans (2.34 mm vs 2.78 mm), and its single-image fitting pipeline predicts meshes more accurately than the previous fitting approach. The result matters because expressive, controllable human models underpin avatars, animation, and image-based human digitization.

Core claim

The paper proposes replacing the vertex-centric paradigm, where joint centers are regressed from the customized surface, with a skeleton-grounded representation. In ATLAS, joint locations are a function only of the skeletal basis and pose, not of the surface shape basis: the external shape basis modifies soft tissue in the rest pose, then a separate skeletal basis of 76 controllable attributes scales and poses the mesh through linear blend skinning. The paper also introduces sparse, non-linear pose correctives: each joint group is processed by a small MLP and then mapped to vertex offsets through a geodesically initialized, L1-regularized sparse mask, so deformations stay local around actuat

What carries the argument

The load-bearing object is the decoupled parameterization: a fixed template skeleton, a linear surface-shape basis that edits soft tissue in the rest pose, and a separate skeletal basis of 76 attributes (15 body-part scales and 61 bone-length offsets) that scales and poses the mesh via linear blend skinning. The argument is carried by the identity that makes joint positions independent of the surface shape components. A secondary mechanism is the sparse, non-linear pose corrective, a per-joint-group MLP followed by a geodesic-initialized, L1-regularized sparse mask, which localizes pose-dependent deformations to nearby vertices.

Load-bearing premise

The decoupling holds only if the two-stage registration recovers the true internal skeleton from keypoints measured on the outer skin; if flesh thickness leaks into those keypoint fits, skeleton and shape are not fully independent.

What would settle it

Take a set of subjects, measure their true bone lengths with MRI, CT, or dual-energy X-ray, and compare against ATLAS bone lengths recovered from the scanning protocol used for training. If the ATLAS skeleton changes systematically with body-fat fraction or soft-tissue thickness at fixed height, the decoupling has failed; stable bone lengths across weight changes would confirm it.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • Body editing becomes deterministic: a single skeletal attribute changes shoulder width or arm length, and a surface component changes weight, without the two interfering.
  • Keypoint-based fitting no longer needs to distort soft tissue, because skeleton parameters are optimized against keypoints while shape parameters are optimized against silhouettes and depth.
  • Sparse non-linear pose correctives improve fitting over both sparse linear and dense non-linear alternatives, with gains concentrated around elbows, knees, and shoulders.
  • The 115k-vertex high-resolution mesh skins in about 5.4 ms on an A100, and the model supports standard lower-resolution topologies for compatibility with existing pipelines.

Where Pith is reading between the lines

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

  • The clean separation suggests a downstream use the paper does not develop: varying bone length and soft tissue independently in ergonomic or medical simulation, where stature and body composition are separate inputs.
  • A direct test of the decoupling would compare ATLAS's recovered bone lengths to MRI or CT bone geometry for the same subjects; if the skeleton estimates drift with body-mass index, the independence is only as good as the keypoint-based registration.
  • The sparse non-linear corrective recipe should transfer to other articulated structures such as hands or quadruped bodies, since it relies only on local joint neighborhoods and geodesic masks.
  • Because the skeletal basis is only 16 components and the surface basis 128, the model offers a low-dimensional factored latent space that generative models or neural avatars could condition on separately for skeleton and tissue.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. The paper introduces ATLAS, a parametric human body model that explicitly separates a surface shape basis (128 components) from a skeletal basis (16 components over 76 attributes: 15 part scales and 61 bone-length offsets). Surface vertices are first customized in the A-pose while the skeleton stays fixed; the mesh is then scaled and posed via LBS, so joint locations depend only on skeleton and pose parameters (Eq. 2). Pose-dependent deformations are modeled by a sparse, non-linear corrective function (Eqs. 3–4). The model is trained on 600k high-resolution scans plus CAESAR/SizeUSA, and evaluated by fitting to 3DBodyTex and a held-out Goliath-Test, reporting lower vertex errors than SMPL-X, STAR, and SUPR. A single-image fitting pipeline and a runtime comparison are also presented.

Significance. If the decoupling is validated, ATLAS would be a meaningful advance: it offers large-scale training data, a high-resolution mesh (115k vertices), fine-grained skeletal control, and sparse non-linear pose correctives. The architectural formulation is clear and the supplementary material provides detailed training and LBS details. However, the central claim of independence between skeleton and soft tissue needs stronger support, both at the level of what the learned skeletal basis represents and in the statistical reliability of the empirical comparisons.

major comments (3)
  1. [§3.1–§3.2, §4.1, Eq. (2)] The abstract and §3.1 claim that keypoint fitting is 'independent of external soft-tissue characteristics,' but the decoupling is only partial. The 76 skeletal attributes include 15 scale parameters that change body-part size (Sec. 3.2), and Eq. (2) applies them through LBS to surface vertices, so β_k can alter soft tissue as well as joints. Additionally, the skeleton-only registrations in Sec. 4.1 are regularized by triangulated keypoints from the outer surface; the resulting skeletal basis may absorb fat/muscle thickness rather than bone geometry. Please validate the skeletal space with an independent anatomical source or show that keypoint-driven β_k changes do not systematically affect body-mass-related surface attributes. Without this, the core decoupling claim is not established.
  2. [§4.2, Fig. 5, Tables 3–4] The main empirical claims are single point estimates without error bars or significance tests. The headline Goliath-Test result (2.34 mm vs 2.78 mm), the 3DBodyTex component curves, the ablation in Table 4, and the monocular fitting results in Table 3 all need per-subject variance and paired statistical tests to support the claim that ATLAS outperforms baselines. Please report mean±std, confidence intervals, and appropriate tests (e.g., Wilcoxon signed-rank) for key comparisons.
  3. [§4.3 (Linear vs Non-Linear Pose Correctives)] The comparison of non-linear vs linear pose correctives reports only a single error reduction (1.82→1.61 mm) on the SMPL dataset. No error bars, number of test sequences, or per-joint error analysis are given, and the non-linear model has strictly more capacity. To support the specific claim that the sparsity mechanism is beneficial, please report statistical significance, parameter-matched baselines, or quantitative sparsity measures (e.g., number of active vertices per joint).
minor comments (5)
  1. [§3.3, Eq. (3)] Typo: 'immediate immediate' in the sentence introducing the local neighbor set n(j).
  2. [Fig. 5] Please clarify how the 'number of fitting components' is counted for each model; for ATLAS it appears to be shape+scale, while for baselines it is shape only. Also provide the absolute error values behind the '21.6% lower' claim in a table.
  3. [Table 2] Clarify whether the reported runtime includes evaluation of the pose-corrective network and whether it is averaged over a standard pose sequence; also state the exact GPU model beyond 'A100'.
  4. [§3.4] Please describe how Sapiens relative depth is normalized and how the rendered depth is aligned to it; this is important for reproducibility of E_depth.
  5. [Supplementary Table 4] This ablation is only in the supplement; consider moving it to the main text and adding error bars, since it directly supports the necessity of both shape and skeleton parameters.

Circularity Check

0 steps flagged

No significant circularity: ATLAS's derivation is self-contained; the decoupled skeleton/shape model is defined by construction and evaluated on held-out/external data.

full rationale

The paper's central derivation chain is not circular. ATLAS defines an architectural decoupling in Eq. (2): joint locations are functions of skeletal components βk and pose θ only, while surface shape βs affects the unposed, un-scaled mesh before LBS. This is an explicit model definition, not a result derived from a re-used input. The model is learned from 600k scans plus external datasets (CAESAR, SizeUSA) via autoencoders over surface vertices and skeletal attributes, and quantitative evaluation is performed on the held-out Goliath-Test and the external 3DBodyTex benchmark. The two-stage registration in Section 4.1 first fits skeletal parameters using triangulated keypoints and then fits surface shape; this is a data-generation procedure that could affect anatomical validity, but it does not make any reported prediction equivalent to a fitted input by construction. The only self-citation of note is Sapiens [24], used as a component in the monocular fitting pipeline; that citation is not load-bearing for the body model's decoupling claim. The reviewer concern about keypoints reflecting soft-tissue thickness is a potential confound in the training data, not a circularity of the derivation.

Axiom & Free-Parameter Ledger

3 free parameters · 4 axioms · 1 invented entities

The central modeling cost of ATLAS is the hand-designed skeleton parameterization and the two-stage registration that separates skeleton from surface. The model depends on the assumption that keypoint-based skeleton fits are independent of soft tissue, and on the scale and representativeness of the proprietary Goliath dataset.

free parameters (3)
  • number of surface shape components = 128
    Chosen by hand; the paper says fitting error plateaued beyond these components (Section E.2).
  • number of skeleton components = 16
    Chosen by hand; same rationale as shape components (Section E.2).
  • pose-corrective feature dimension = 24
    Chosen by hand; architecture hyperparameter (Section 4.1).
axioms (4)
  • domain assumption Triangulated keypoints derived from surface-scanned subjects provide unbiased estimates of internal skeletal joint locations.
    Section 4.1 Implementation Details: the skeleton-only fits use triangulated keypoints to regularize joints; if keypoints are influenced by soft tissue, the decoupling is imperfect.
  • domain assumption The Goliath dataset of 600k scans from 130 subjects is representative of the human shape and pose distribution for training a generalizable body model.
    Claims generalization to unseen subjects (Section 4.2) rely on this; 130 subjects is a limited demographic sample.
  • domain assumption The artist-designed skeleton with sub-joints and skin weights is anatomically consistent.
    Section E.1.2 describes the joint structure designed by sculpting artists; if the skeleton is not anatomically accurate, the skeletal attributes and pose correctives are built on an incorrect rig.
  • standard math Linear autoencoders with ordered dropout preserve a component hierarchy that captures meaningful shape/skeleton variation.
    Section E.2 uses sparse autoencoders initialized with PCA; standard technique.
invented entities (1)
  • 76 controllable skeletal attributes (15 scales, 61 bone-length offsets) no independent evidence
    purpose: Direct manipulation of body-part sizes and bone lengths independently of surface shape.
    These are new model parameters introduced by ATLAS; their utility is demonstrated only in this paper's qualitative and quantitative results, with no external validation.

reviewed 2026-08-05 · how reviews work

0 comments
Cite this review

Pith. "Pith review of ATLAS: Decoupling Skeletal and Shape Parameters for Expressive Parametric Human Modeling." pith.science (2026). https://pith.science/paper/UXRVQ6LW

@misc{pith2026250815767,
  author       = {Pith},
  title        = {Pith review of: ATLAS: Decoupling Skeletal and Shape Parameters for Expressive Parametric Human Modeling},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/UXRVQ6LW}},
  note         = {Machine review of arXiv:2508.15767}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Parametric body models offer expressive 3D representation of humans across a wide range of poses, shapes, and facial expressions, typically derived by learning a basis over registered 3D meshes. However, existing human mesh modeling approaches struggle to capture detailed variations across diverse body poses and shapes, largely due to limited training data diversity and restrictive modeling assumptions. Moreover, the common paradigm first optimizes the external body surface using a linear basis, then regresses internal skeletal joints from surface vertices. This approach introduces problematic dependencies between internal skeleton and outer soft tissue, limiting direct control over body height and bone lengths. To address these issues, we present ATLAS, a high-fidelity body model learned from 600k high-resolution scans captured using 240 synchronized cameras. Unlike previous methods, we explicitly decouple the shape and skeleton bases by grounding our mesh representation in the human skeleton. This decoupling enables enhanced shape expressivity, fine-grained customization of body attributes, and keypoint fitting independent of external soft-tissue characteristics. ATLAS outperforms existing methods by fitting unseen subjects in diverse poses more accurately, and quantitative evaluations show that our non-linear pose correctives more effectively capture complex poses compared to linear models.

discussion (0)

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

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Direct Clinical Joint Angle Extraction from Parametric Body Model Rotation Matrices

    cs.CV 2026-07 conditional novelty 6.0

    A swing-twist decomposition plus a per-body-model calibration table converts body-model rotation matrices into clinical joint angles at 4.50° MAE on OpenCap LabValidation.

Reference graph

Works this paper leans on

65 extracted references · 57 canonical work pages · cited by 1 Pith paper · 1 internal anchor

  1. [1]

    SizeUSA dataset.https://www.tc2.com/size- usa.html, 2017. 5

    work page 2017

  2. [2]

    Articulated body deformation from range scan data.ACM Transactions on Graphics, (Proc

    Brett Allen, Brian Curless, and Zoran Popovi ´c. Articulated body deformation from range scan data.ACM Transactions on Graphics, (Proc. SIGGRAPH), 21(3):612–619, 2002. 3

    work page 2002

  3. [3]

    Learning a correlated model of identity and pose- dependent body shape variation for real-time synthesis

    Brett Allen, Brian Curless, Zoran Popovi ´c, and Aaron Hertz- mann. Learning a correlated model of identity and pose- dependent body shape variation for real-time synthesis. In Proceedings of the 2006 ACM SIGGRAPH/Eurographics symposium on Computer animation, pages 147–156. Cite- seer, 2006. 3

    work page 2006

  4. [4]

    Anguelov, P

    D. Anguelov, P. Srinivasan, D. Koller, S. Thrun, J. Rodgers, and J. Davis. SCAPE: Shape Completion and Animation of PEople.ACM TOG, 24(3):408–416, 2005. 3

    work page 2005

  5. [5]

    Scape: shape completion and animation of people

    Dragomir Anguelov, Praveen Srinivasan, Daphne Koller, Se- bastian Thrun, Jim Rodgers, and James Davis. Scape: shape completion and animation of people. InACM SIGGRAPH 2005 Papers, pages 408–416. 2005. 1

    work page 2005

  6. [6]

    Behave: Dataset and method for tracking human object in- teractions

    Bharat Lal Bhatnagar, Xianghui Xie, Ilya A Petrov, Cristian Sminchisescu, Christian Theobalt, and Gerard Pons-Moll. Behave: Dataset and method for tracking human object in- teractions. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 15935– 15946, 2022. 1

    work page 2022

  7. [7]

    Federica Bogo, Angjoo Kanazawa, Christoph Lassner, Peter Gehler, Javier Romero, and Michael J. Black. Keep it SMPL: Automatic estimation of 3D human pose and shape from a single image. InComputer Vision – ECCV 2016. Springer International Publishing, 2016. 5

    work page 2016

  8. [8]

    Facewarehouse: A 3d facial expression database for visual computing.IEEE Transactions on Visualization and Computer Graphics, 20(3):413–425, 2014

    Chen Cao, Yanlin Weng, Shun Zhou, Yiying Tong, and Kun Zhou. Facewarehouse: A 3d facial expression database for visual computing.IEEE Transactions on Visualization and Computer Graphics, 20(3):413–425, 2014. 3

    work page 2014

  9. [9]

    Hico: A benchmark for recognizing human-object interactions in images

    Yu-Wei Chao, Zhan Wang, Yugeng He, Jiaxuan Wang, and Jia Deng. Hico: A benchmark for recognizing human-object interactions in images. InProceedings of the IEEE inter- national conference on computer vision, pages 1017–1025,

    work page

  10. [10]

    Tensor- based human body modeling

    Yinpeng Chen, Zicheng Liu, and Zhengyou Zhang. Tensor- based human body modeling. InProceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pages 105–112, 2013. 3

    work page 2013

  11. [11]

    Dna-rendering: A diverse neural actor repository for high-fidelity human-centric rendering

    Wei Cheng, Ruixiang Chen, Siming Fan, Wanqi Yin, Keyu Chen, Zhongang Cai, Jingbo Wang, Yang Gao, Zhengming Yu, Zhengyu Lin, et al. Dna-rendering: A diverse neural actor repository for high-fidelity human-centric rendering. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 19982–19993, 2023. 1

    2023

  12. [12]

    Anatomy transfer.ACM Transactions on Graphics (TOG), 32:1 – 8, 2013

    Ali-Hamadi Dicko, Tiantian Liu, Benjamin Gilles, Ladislav Kavan, Franc ¸ois Faure, Olivier Palombi, and Marie-Paule Cani. Anatomy transfer.ACM Transactions on Graphics (TOG), 32:1 – 8, 2013. 3

    work page 2013

  13. [13]

    Lie bodies: A mani- fold representation of 3d human shape

    Oren Freifeld and Michael J Black. Lie bodies: A mani- fold representation of 3d human shape. InComputer Vision– ECCV 2012: 12th European Conference on Computer Vi- sion, Florence, Italy, October 7-13, 2012, Proceedings, Part I 12, pages 1–14. Springer, 2012. 3

    work page 2012

  14. [14]

    Statistical methods for tomographic image restoration.Bull

    Stuart Geman. Statistical methods for tomographic image restoration.Bull. Internat. Statist. Inst., 52:5–21, 1987. 5

    work page 1987

  15. [15]

    Cre- ating and animating subject-specific anatomical models

    Benjamin Gilles, Lionel Reveret, and Dinesh K Pai. Cre- ating and animating subject-specific anatomical models. In Computer Graphics Forum, pages 2340–2351. Wiley Online Library, 2010. 3

    work page 2010

  16. [16]

    A statistical model of human pose and body shape

    Nils Hasler, Carsten Stoll, Martin Sunkel, Bodo Rosenhahn, and H-P Seidel. A statistical model of human pose and body shape. InComputer graphics forum, pages 337–346. Wiley Online Library, 2009. 3

    work page 2009

  17. [17]

    Learning skeletons for shape and pose

    Nils Hasler, Thorsten Thorm ¨ahlen, Bodo Rosenhahn, and Hans-Peter Seidel. Learning skeletons for shape and pose. InProceedings of the 2010 ACM SIGGRAPH symposium on Interactive 3D Graphics and Games, pages 23–30, 2010. 3

    work page 2010

  18. [18]

    Coregistration: Simultaneous alignment and modeling of articulated 3d shape

    David A Hirshberg, Matthew Loper, Eric Rachlin, and Michael J Black. Coregistration: Simultaneous alignment and modeling of articulated 3d shape. InComputer Vision– ECCV 2012: 12th European Conference on Computer Vi- sion, Florence, Italy, October 7-13, 2012, Proceedings, Part VI 12, pages 242–255. Springer, 2012. 3

    work page 2012

  19. [19]

    Total cap- ture: A 3D deformation model for tracking faces, hands, and bodies

    Hanbyul Joo, Tomas Simon, and Yaser Sheikh. Total cap- ture: A 3D deformation model for tracking faces, hands, and bodies. InCVPR, pages 8320–8329, 2018. 3

    work page 2018

  20. [20]

    Reconstructing per- sonalized anatomical models for physics-based body anima- tion.ACM Transactions on Graphics (TOG), 35(6):1–13,

    Petr Kadle ˇcek, Alexandru-Eugen Ichim, Tiantian Liu, Jaroslav Kˇriv´anek, and Ladislav Kavan. Reconstructing per- sonalized anatomical models for physics-based body anima- tion.ACM Transactions on Graphics (TOG), 35(6):1–13,

    work page

  21. [21]

    Skinning with dual quaternions

    Ladislav Kavan, Steven Collins, Ji ˇr´ı ˇZ´ara, and Carol O’Sullivan. Skinning with dual quaternions. InProceed- ings of the 2007 symposium on Interactive 3D graphics and games, pages 39–46, 2007. 2, 3, 4

    work page 2007

  22. [22]

    Osso: Obtaining skeletal shape from outside

    Marilyn Keller, Silvia Zuffi, Michael J Black, and Sergi Pu- jades. Osso: Obtaining skeletal shape from outside. InPro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 20492–20501, 2022. 2, 3

    work page 2022

  23. [23]

    From skin to skeleton: Towards biomechanically accurate 3d dig- ital humans.ACM Transactions on Graphics (TOG), 42(6): 1–12, 2023

    Marilyn Keller, Keenon Werling, Soyong Shin, Scott Delp, Sergi Pujades, C Karen Liu, and Michael J Black. From skin to skeleton: Towards biomechanically accurate 3d dig- ital humans.ACM Transactions on Graphics (TOG), 42(6): 1–12, 2023. 2, 3

    2023

  24. [24]

    Sapiens: Foundation for human vision mod- els

    Rawal Khirodkar, Timur Bagautdinov, Julieta Martinez, Su Zhaoen, Austin James, Peter Selednik, Stuart Anderson, and Shunsuke Saito. Sapiens: Foundation for human vision mod- els. InEuropean Conference on Computer Vision, pages 198–213. Springer, 2025. 2, 5

    work page 2025

  25. [25]

    Auto-encoding varia- tional bayes.arXiv preprint arXiv:1312.6114, 2013

    Diederik P Kingma and Max Welling. Auto-encoding varia- tional bayes.arXiv preprint arXiv:1312.6114, 2013. 2

    Pith/arXiv arXiv 2013

  26. [26]

    Eigenskin: real time large deformation character skinning in hardware

    Paul G Kry, Doug L James, and Dinesh K Pai. Eigenskin: real time large deformation character skinning in hardware. InProceedings of the 2002 ACM SIGGRAPH/Eurographics symposium on Computer animation, pages 153–159. ACM,

    work page 2002

  27. [27]

    Modeling de- formable human hands from medical images

    Tsuneya Kurihara and Natsuki Miyata. Modeling de- formable human hands from medical images. InProceed- ings of the 2004 ACM SIGGRAPH/Eurographics symposium on Computer animation, pages 355–363. Eurographics As- sociation, 2004. 3

    work page 2004

  28. [28]

    Comprehensive biomechanical modeling and simulation of the upper body.ACM Transactions on Graphics (TOG), 28 (4):1–17, 2009

    Sung-Hee Lee, Eftychios Sifakis, and Demetri Terzopoulos. Comprehensive biomechanical modeling and simulation of the upper body.ACM Transactions on Graphics (TOG), 28 (4):1–17, 2009. 3

    work page 2009

  29. [29]

    J. P. Lewis, Matt Cordner, and Nickson Fong. Pose space deformation: A unified approach to shape interpolation and skeleton-driven deformation. InProceedings of the 27th Annual Conference on Computer Graphics and Interactive Techniques, pages 165–172, New York, NY , USA, 2000. ACM Press/Addison-Wesley Publishing Co. 3

    work page 2000

  30. [30]

    Tianye Li, Timo Bolkart, Michael. J. Black, Hao Li, and Javier Romero. Learning a model of facial shape and ex- pression from 4D scans.ACM Transactions on Graphics, (Proc. SIGGRAPH Asia), 36(6), 2017. 5

    work page 2017

  31. [31]

    Rich human feedback for text-to-image generation

    Youwei Liang, Junfeng He, Gang Li, Peizhao Li, Arseniy Klimovskiy, Nicholas Carolan, Jiao Sun, Jordi Pont-Tuset, Sarah Young, Feng Yang, et al. Rich human feedback for text-to-image generation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 19401–19411, 2024. 1

    work page 2024

  32. [32]

    Neural vol- umes: Learning dynamic renderable volumes from images

    Stephen Lombardi, Tomas Simon, Jason Saragih, Gabriel Schwartz, Andreas Lehrmann, and Yaser Sheikh. Neural vol- umes: Learning dynamic renderable volumes from images. arXiv preprint arXiv:1906.07751, 2019. 1

    Pith/arXiv arXiv 1906

  33. [33]

    Matthew Loper, Naureen Mahmood, Javier Romero, Gerard Pons-Moll, and Michael J. Black. SMPL: A skinned multi- person linear model.ACM Transactions on Graphics, (Proc. SIGGRAPH Asia), 34(6):248:1–248:16, 2015. 1, 2, 3, 4, 5, 6, 7, 12

    work page 2015

  34. [34]

    Troje, Ger- ard Pons-Moll, and Michael J

    Naureen Mahmood, Nima Ghorbani, Nikolaus F. Troje, Ger- ard Pons-Moll, and Michael J. Black. AMASS: Archive of motion capture as surface shapes. InICCV, pages 5442– 5451, 2019. 3

    work page 2019

  35. [35]

    Sparse autoencoder.CS294A Lecture notes, 72(2011):1–19, 2011

    Andrew Ng et al. Sparse autoencoder.CS294A Lecture notes, 72(2011):1–19, 2011. 14

    work page 2011

  36. [36]

    Efficient trajectory optimization for curved running using a 3d musculoskeletal model with im- plicit dynamics.Scientific reports, 10(1):17655, 2020

    Marlies Nitschke, Eva Dorschky, Dieter Heinrich, Heiko Schlarb, Bjoern M Eskofier, Anne D Koelewijn, and An- tonie J van den Bogert. Efficient trajectory optimization for curved running using a 3d musculoskeletal model with im- plicit dynamics.Scientific reports, 10(1):17655, 2020. 3

    work page 2020

  37. [37]

    Supr: A sparse unified part-based human representation

    Ahmed AA Osman, Timo Bolkart, Dimitrios Tzionas, and Michael J Black. Supr: A sparse unified part-based human representation. InEuropean Conference on Computer Vi- sion, pages 568–585. Springer, 2022. 2, 3, 4, 6

    work page 2022

  38. [38]

    Ahmed A. A. Osman, Timo Bolkart, and Michael J. Black. STAR: Sparse trained articulated human body regressor. In ECCV, pages 598–613, 2020. 2, 3, 4, 6

    work page 2020

  39. [39]

    Georgios Pavlakos, Vasileios Choutas, Nima Ghorbani, Timo Bolkart, Ahmed A. A. Osman, Dimitrios Tzionas, and Michael J. Black. Expressive body capture: 3d hands, face, and body from a single image. InProceedings IEEE Conf. on Computer Vision and Pattern Recognition (CVPR), 2019. 1, 2, 3, 4, 5, 6, 7, 8, 14

    work page 2019

  40. [40]

    Charactergen: Efficient 3d character generation from single images with multi-view pose canon- icalization.ACM Transactions on Graphics (TOG), 43(4): 1–13, 2024

    Hao-Yang Peng, Jia-Peng Zhang, Meng-Hao Guo, Yan-Pei Cao, and Shi-Min Hu. Charactergen: Efficient 3d character generation from single images with multi-view pose canon- icalization.ACM Transactions on Graphics (TOG), 43(4): 1–13, 2024. 1

    work page 2024

  41. [41]

    Neural body: Implicit neural representations with structured latent codes for novel view synthesis of dynamic humans

    Sida Peng, Yuanqing Zhang, Yinghao Xu, Qianqian Wang, Qing Shuai, Hujun Bao, and Xiaowei Zhou. Neural body: Implicit neural representations with structured latent codes for novel view synthesis of dynamic humans. InCVPR,

  42. [42]

    Rasterized edge gradients: Handling discontinuities differentiably.ArXiv, abs/2405.02508, 2024

    Stanislav Pidhorskyi, Tomas Simon, Gabriel Schwartz, He Wen, Yaser Sheikh, and Jason Saragih. Rasterized edge gradients: Handling discontinuities differentiably.ArXiv, abs/2405.02508, 2024. 5

    Pith/arXiv arXiv 2024

  43. [43]

    Dyna: A model of dynamic human shape in motion.ACM Transactions on Graphics (TOG), 34(4):1–14,

    Gerard Pons-Moll, Javier Romero, Naureen Mahmood, and Michael J Black. Dyna: A model of dynamic human shape in motion.ACM Transactions on Graphics (TOG), 34(4):1–14,

    work page

  44. [44]

    Full-body musculoskeletal model for muscle-driven simula- tion of human gait.IEEE transactions on biomedical engi- neering, 63(10):2068–2079, 2016

    Apoorva Rajagopal, Christopher L Dembia, Matthew S De- Mers, Denny D Delp, Jennifer L Hicks, and Scott L Delp. Full-body musculoskeletal model for muscle-driven simula- tion of human gait.IEEE transactions on biomedical engi- neering, 63(10):2068–2079, 2016. 3

    work page 2068

  45. [45]

    Make-a-character: High quality text- to-3d character generation within minutes.arXiv preprint arXiv:2312.15430, 2023

    Jianqiang Ren, Chao He, Lin Liu, Jiahao Chen, Yutong Wang, Yafei Song, Jianfang Li, Tangli Xue, Siqi Hu, Tao Chen, et al. Make-a-character: High quality text- to-3d character generation within minutes.arXiv preprint arXiv:2312.15430, 2023. 1

    Pith/arXiv arXiv 2023

  46. [46]

    Real- time weighted pose-space deformation on the gpu

    Taehyun Rhee, John P Lewis, and Ulrich Neumann. Real- time weighted pose-space deformation on the gpu. InCom- puter Graphics Forum, pages 439–448. Wiley Online Li- brary, 2006. 3

    work page 2006

  47. [47]

    Robinette, Sherri Blackwell, Hein Daanen, Mark Boehmer, Scott Fleming, Tina Brill, David Hoeferlin, and Dennis Burnsides

    Kathleen M. Robinette, Sherri Blackwell, Hein Daanen, Mark Boehmer, Scott Fleming, Tina Brill, David Hoeferlin, and Dennis Burnsides. Civilian American and European Sur- face Anthropometry Resource (CAESAR) final report. Tech- nical Report AFRL-HE-WP-TR-2002-0169, US Air Force Research Laboratory, 2002. 5

    work page 2002

  48. [48]

    Em- bodied hands: Modeling and capturing hands and bodies to- gether.ACM TOG, 36(6):245:1–245:17, 2017

    Javier Romero, Dimitrios Tzionas, and Michael J Black. Em- bodied hands: Modeling and capturing hands and bodies to- gether.ACM TOG, 36(6):245:1–245:17, 2017. 3

    work page 2017

  49. [49]

    3dbodytex: Textured 3d body dataset

    Alexandre Saint, Eman Ahmed, Abd El Rahman Shabayek, Kseniya Cherenkova, Gleb Gusev, Djamila Aouada, and Bjorn Ottersten. 3dbodytex: Textured 3d body dataset. In 2018 International Conference on 3D Vision (3DV), pages 495–504, 2018. 2, 6

    work page 2018

  50. [50]

    Compu- tational bodybuilding: Anatomically-based modeling of hu- man bodies.ACM Transactions on Graphics (TOG), 34(4): 1–12, 2015

    Shunsuke Saito, Zi-Ye Zhou, and Ladislav Kavan. Compu- tational bodybuilding: Anatomically-based modeling of hu- man bodies.ACM Transactions on Graphics (TOG), 34(4): 1–12, 2015. 3

    work page 2015

  51. [51]

    Pifuhd: Multi-level pixel-aligned implicit function for high-resolution 3d human digitization

    Shunsuke Saito, Tomas Simon, Jason Saragih, and Hanbyul Joo. Pifuhd: Multi-level pixel-aligned implicit function for high-resolution 3d human digitization. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 84–93, 2020. 1

    work page 2020

  52. [52]

    A biomechanical model of the scapulothoracic joint to accurately capture scapular kinematics during shoulder movements.PloS one, 11(1):e0141028, 2016

    Ajay Seth, Ricardo Matias, Ant ´onio P Veloso, and Scott L Delp. A biomechanical model of the scapulothoracic joint to accurately capture scapular kinematics during shoulder movements.PloS one, 11(1):e0141028, 2016. 3

    work page 2016

  53. [53]

    Boss: Bones, organs and skin shape model

    Karthik Shetty, Annette Birkhold, Srikrishna Jaganathan, Norbert Strobel, Bernhard Egger, Markus Kowarschik, and Andreas Maier. Boss: Bones, organs and skin shape model. Computers in Biology and Medicine, 165:107383, 2023. 2, 3

    work page 2023

  54. [54]

    Blsm: A bone-level skinned model of the human mesh

    Haoyang Wang, Riza Alp G ¨uler, Iasonas Kokkinos, George Papandreou, and Stefanos Zafeiriou. Blsm: A bone-level skinned model of the human mesh. InComputer Vision– ECCV 2020: 16th European Conference, Glasgow, UK, August 23–28, 2020, Proceedings, Part V 16, pages 1–17. Springer, 2020. 1, 2, 3

    work page 2020

  55. [55]

    A Review of Human-Object Interaction Detection

    Yuxiao Wang, Qiwei Xiong, Yu Lei, Weiying Xue, Qi Liu, and Zhenao Wei. A review of human-object interaction de- tection.arXiv preprint arXiv:2408.10641, 2024. 1

    work page internal anchor Pith review Pith/arXiv arXiv 2024

  56. [56]

    Hu- mannerf: Free-viewpoint rendering of moving people from monocular video

    Chung-Yi Weng, Brian Curless, Pratul P Srinivasan, Jonathan T Barron, and Ira Kemelmacher-Shlizerman. Hu- mannerf: Free-viewpoint rendering of moving people from monocular video. InProceedings of the IEEE/CVF con- ference on computer vision and pattern Recognition, pages 16210–16220, 2022. 1

    work page 2022

  57. [57]

    Keenon Werling, Michael Raitor, Jon Stingel, Jennifer L Hicks, Steve Collins, Scott L Delp, and C Karen Liu. Rapid bilevel optimization to concurrently solve musculoskeletal scaling, marker registration, and inverse kinematic problems for human motion reconstruction.bioRxiv, pages 2022–08,

    work page 2022

  58. [58]

    Icon: Implicit clothed humans obtained from nor- mals

    Yuliang Xiu, Jinlong Yang, Dimitrios Tzionas, and Michael J Black. Icon: Implicit clothed humans obtained from nor- mals. In2022 IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition (CVPR), pages 13286–13296. IEEE, 2022. 1

    work page 2022

  59. [59]

    GHUM & GHUML: Generative 3D human shape and articulated pose models

    Hongyi Xu, Eduard Gabriel Bazavan, Andrei Zanfir, William T Freeman, Rahul Sukthankar, and Cristian Smin- chisescu. GHUM & GHUML: Generative 3D human shape and articulated pose models. InCVPR, pages 6184–6193,

    work page

  60. [60]

    FaceScape: a large- scale high quality 3D face dataset and detailed riggable 3D face prediction

    Haotian Yang, Hao Zhu, Yanru Wang, Mingkai Huang, Qiu Shen, Ruigang Yang, and Xun Cao. FaceScape: a large- scale high quality 3D face dataset and detailed riggable 3D face prediction. InCVPR, pages 601–610, 2020. 1

    work page 2020

  61. [61]

    Hi4d: 4d instance seg- mentation of close human interaction

    Yifei Yin, Chen Guo, Manuel Kaufmann, Juan Jose Zarate, Jie Song, and Otmar Hilliges. Hi4d: 4d instance seg- mentation of close human interaction. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 17016–17027, 2023. 1

    work page 2023

  62. [62]

    Structured local radiance fields for human avatar modeling

    Zerong Zheng, Han Huang, Tao Yu, Hongwen Zhang, Yan- dong Guo, and Yebin Liu. Structured local radiance fields for human avatar modeling. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 15893–15903, 2022. 1

    work page 2022

  63. [63]

    On the continuity of rotation representations in neural networks

    Yi Zhou, Connelly Barnes, Jingwan Lu, Jimei Yang, and Hao Li. On the continuity of rotation representations in neural networks. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 5745–5753,

  64. [64]

    Adaptable anatomical models for realistic bone motion reconstruction

    Lifeng Zhu, Xiaoyan Hu, and Ladislav Kavan. Adaptable anatomical models for realistic bone motion reconstruction. InComputer Graphics Forum, pages 459–471. Wiley Online Library, 2015. 3 A. Supplementary Overview In the supplementary video, we present video results of fit- ting ATLAS to high-fidelity 3D scans, demonstrate control- lability of skeletal attr...

    work page 2015

  65. [65]

    Of particular note is ATLAS’s ease at capturing undersized subjects such as children

    Our fitting procedure complements ATLAS by yield- ing shape, scale, pose, and expression parameters from 2D RGB images in the wild. Of particular note is ATLAS’s ease at capturing undersized subjects such as children. By explicitly modeling the size of each skeletal part, ATLAS naturally predicts realistic shapes for children, accounting for their relativ...

    work page

This paper was first reviewed by deepseek-v4-flash on August 5, 2026.