pith. sign in

arxiv: 2605.18645 · v1 · pith:BZG3JCWJnew · submitted 2026-05-18 · 💻 cs.CV

Articulation in Prime: Primitive-Based Articulated Object Understanding from a Single Casual Video

Arslan Artykov , Tom Ravaud , Nicol\'as Violante-Grezzi , Vincent Lepetit This is my paper

Pith reviewed 2026-05-20 11:57 UTC · model grok-4.3

classification 💻 cs.CV
keywords articulated objectsprimitive fittingmonocular videokinematics recoverypart segmentationrevolute jointsprismatic jointsoptimization framework
0
0 comments X

The pith

Geometric primitives recover articulated kinematics from a single casual video by jointly optimizing parts and joints.

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

The paper establishes a category-agnostic optimization framework that reframes articulated object understanding as a primitive-fitting task. Geometric primitives act as stable proxies for object parts, sidestepping the instability of point tracks under occlusions or fast motion. These primitives are grouped into coherent parts connected by revolute or prismatic joints. The approach simultaneously solves for part segmentation and joint parameters while incorporating a visibility-aware procedure to manage partial views. This yields accurate 3D kinematics recovery from everyday monocular videos and shows stronger results than prior methods on new benchmarks built around heavy occlusions and camera movement.

Core claim

We propose a category-agnostic optimization framework that treats articulated object understanding as a primitive-fitting problem. Geometric primitives serve as a proxy representation that avoids the pitfalls of unstable point tracks; a novel mechanism organizes them into coherent parts constrained by revolute and prismatic joints. Our formulation jointly optimizes part segmentation and joint parameters, recovering complex kinematics from a single casually captured video. A visibility-aware procedure handles partial observations and occlusions inherent to real-world data.

What carries the argument

Geometric primitives fitted across video frames and grouped into parts via revolute and prismatic joint constraints; the mechanism performs the work of providing a stable proxy that supports simultaneous optimization of segmentation and kinematics parameters.

If this is right

  • Recovers complex 3D kinematics from monocular video without long-term point tracking or wide-baseline matching.
  • Operates without category-specific training data or specialized capture rigs.
  • Handles severe occlusions and rapid ego-motion through the visibility-aware procedure.
  • Outperforms prior methods on the AiP-synth and AiP-real benchmarks that emphasize challenging real-world conditions.

Where Pith is reading between the lines

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

  • The same primitive proxy idea could be tested on videos containing spherical or screw joints to check the limits of the current revolute-prismatic constraint set.
  • Combining the optimization with learned priors might improve initialization speed while retaining the stability of explicit fitting.
  • The approach could extend to multi-object scenes if the organization mechanism is augmented with inter-object separation terms.

Load-bearing premise

Geometric primitives provide a stable enough representation to segment parts and recover joint parameters accurately even when occlusions, rapid camera motion, and weak features make point tracking unreliable.

What would settle it

If joint parameters recovered on the AiP-real benchmark videos deviate substantially from known ground-truth kinematics under documented heavy occlusions, the claim that primitives serve as a reliable proxy would not hold.

Figures

Figures reproduced from arXiv: 2605.18645 by Arslan Artykov, Nicol\'as Violante-Grezzi, Tom Ravaud, Vincent Lepetit.

Figure 1
Figure 1. Figure 1: Overview. Given a sequence of depth maps and camera poses extracted from a video, we backproject the frames into a sequence of partial point clouds in world coordinates. We jointly optimize a set of superquadric primitives and part assignments, where each primitive is softly assigned to a part via a differentiable allocation, and each part carries its own joint parameters (revolute or prismatic) and per-ti… view at source ↗
Figure 2
Figure 2. Figure 2: Qualitative results on the Arti4D dataset. Red, green, and blue arrows denote Ours, Artipoint, and Ground-Truth joint axes, respectively. For prismatic (translational) joints, only the axis direction is evaluated, while the axis location is not considered. Because Artipoint produces a large error in axis location for scene RH078, the green joint falls outside the camera frustum. 4.2 Datasets Video2Articula… view at source ↗
Figure 3
Figure 3. Figure 3: Qualitative results on randomly selected sequences from the AiP-synth dataset. Red arrows denote predicted joint axes. Note how our method retrieves the part segmentations and the joint axes much more accurately and robustly than all the other methods. ’×’ indicates that the method fails on the category. Video results are provided in the video supplementary material. We observe a significant degradation in… view at source ↗
Figure 4
Figure 4. Figure 4: Qualitative results on the real-world sequences from the AiP-real dataset. Red arrows denote predicted joint axes. Note how our method retrieves the part segmentations and the joint axes much more accurately and robustly than all the other methods. ’×’ indicates that the method fails on the category. Video results are provided in the video supplementary material [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Influence of the shape parameters on a unit scale superquadric. The shape parameters are constrained to the interval [0.1, 1.9] so that superquadrics are convex and more representative of shapes encountered in the real world. A.1 Equations The superquadric surface is defined by the implicit equation: " x sx 2/ε2 +  y sy 2/ε2 #ε2/ε1 +  z sz 2/ε1 = 1 (20) or equivalently, by the explicit equation:   … view at source ↗
Figure 6
Figure 6. Figure 6: Representative frames of all the objects used in our AiP-synth dataset. This selection covers a wide range of geometries and kinematics. Some object parts are particularly complex and require the use of multiple primitives. Other objects, such as the globe, exhibit symmetries that make the joint discovery more challenging. C Details on AiP-synth and AiP-real Datasets C.1 AiP-synth The dataset comprises a d… view at source ↗
Figure 7
Figure 7. Figure 7: A synthetic circular camera trajectory used in AiP-synth. The camera turns around the object with random start and end angles, and a random height variation amplitude. In the case of the ping-pong trajectory, the camera sweeps back and forth along the circular path. Circular trajectory. Let t ∈ [0, 1] parametrize the trajectory. The azimuth angle is linearly interpolated between the start and end angles: θ… view at source ↗
Figure 8
Figure 8. Figure 8: shows qualitative results on the remaining sequences of the AiP-synth dataset. Our approach again yields better results than competing methods. Ground Truth Ours Articulate- Anything ReArt Video2 Articulation Bottle Faucet Globe Lighter Scissors Drawer Fridge Oven USB Phone [PITH_FULL_IMAGE:figures/full_fig_p016_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Fitted superquadrics results on sequences from the AiP-synth dataset. Red arrows denote predicted joint axes. Representative Frame Superquadrics Box Chair Sliding Box Globe Excavator [PITH_FULL_IMAGE:figures/full_fig_p017_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Fitted superquadrics results on the sequences from the AiP-real dataset. Red arrows denote predicted joint axes. and AiP-real objects in Figures 9 and 10, respectively. We observe that certain parts automatically aggregate multiple primitives to reconstruct complex object structures beyond the modeling capacity of a single superquadric, such as the lamp base, the chair body or the excavator arm. Despite t… view at source ↗
read the original abstract

Retrieving the 3D kinematics of articulated objects from monocular video is a fundamental challenge in computer vision. Existing methods rely on complex video setups or cues such as long-term point tracking or wide-baseline matching, but are frequently brittle under severe occlusions, rapid camera ego-motion, or weak local features. Learning-based methods, meanwhile, struggle to generalize beyond their training categories. We propose a category-agnostic optimization framework that treats articulated object understanding as a primitive-fitting problem. Geometric primitives serve as a proxy representation that avoids the pitfalls of unstable point tracks; a novel mechanism organizes them into coherent parts constrained by revolute and prismatic joints. Our formulation jointly optimizes part segmentation and joint parameters, recovering complex kinematics from a single casually captured video. A visibility-aware procedure handles partial observations and occlusions inherent to real-world data. We also propose the AiP-synth and AiP-real benchmarks, featuring significant camera motion and heavy occlusions, and outperform existing methods. Project page: https://aartykov.github.io/Articulation-in-Prime/

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 / 2 minor

Summary. The manuscript proposes a category-agnostic optimization framework for recovering 3D kinematics of articulated objects from a single casually captured monocular video. Geometric primitives act as proxy representations for object parts; these are organized into coherent rigid bodies connected by revolute or prismatic joints. A visibility-aware energy jointly optimizes primitive parameters, part segmentation, and joint axes while handling occlusions and ego-motion. New AiP-synth and AiP-real benchmarks are introduced, and the method is reported to outperform prior approaches on them.

Significance. If the optimization is shown to be well-constrained and robust, the primitive-proxy formulation would constitute a useful alternative to point-tracking or learning-based pipelines for articulated-object understanding, especially under severe occlusions and rapid camera motion. The creation of dedicated benchmarks that emphasize these challenging conditions is a concrete contribution that future work can build upon.

major comments (2)
  1. [§4] §4 (Optimization Framework) and the visibility-aware energy: the central claim that primitives simultaneously yield stable segmentation and kinematics rests on the assumption that the energy supplies enough independent constraints to disentangle per-primitive 3D pose/scale, part assignment, and joint-axis parameters while the camera trajectory remains free. The manuscript should provide an explicit analysis or ablation showing that the formulation avoids degenerate solutions (e.g., axis flips or part swaps) under rapid ego-motion; without such analysis the joint-optimization claim remains under-supported.
  2. [Experiments] Experiments section, benchmark tables: the reported outperformance on AiP-synth and AiP-real is presented without error bars, multiple random initializations, or sensitivity analysis. Given the skeptic concern about initialization dependence, the absence of these statistics makes it impossible to judge whether the quantitative gains are reliable or merely reflect favorable starting points.
minor comments (2)
  1. [Abstract] The abstract and method overview use the term “visibility-aware procedure” without a forward reference to the precise equation or subsection that defines the visibility term.
  2. [§3] Notation for primitive parameters (pose, scale, axis) should be collected in a single table or paragraph early in the method section to improve readability.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on our manuscript. We address each major comment below and have revised the manuscript to strengthen the supporting analysis and experimental reporting.

read point-by-point responses

  1. Referee: [§4] §4 (Optimization Framework) and the visibility-aware energy: the central claim that primitives simultaneously yield stable segmentation and kinematics rests on the assumption that the energy supplies enough independent constraints to disentangle per-primitive 3D pose/scale, part assignment, and joint-axis parameters while the camera trajectory remains free. The manuscript should provide an explicit analysis or ablation showing that the formulation avoids degenerate solutions (e.g., axis flips or part swaps) under rapid ego-motion; without such analysis the joint-optimization claim remains under-supported.

    Authors: We appreciate the referee highlighting the need for explicit validation of non-degeneracy. The visibility-aware energy, combined with primitive geometric constraints and revolute/prismatic joint regularization, supplies cross-frame consistency that discourages axis flips and part swaps even under free camera motion. Our AiP benchmarks already contain sequences with rapid ego-motion and heavy occlusions, and the reported results exhibit stable kinematics without observed degeneracies. To make this explicit, we have added a targeted ablation in the revised Section 4.3 that perturbs camera trajectories and initializations, confirming that the recovered axes and part assignments remain consistent. revision: yes

  2. Referee: [Experiments] Experiments section, benchmark tables: the reported outperformance on AiP-synth and AiP-real is presented without error bars, multiple random initializations, or sensitivity analysis. Given the skeptic concern about initialization dependence, the absence of these statistics makes it impossible to judge whether the quantitative gains are reliable or merely reflect favorable starting points.

    Authors: We agree that variability statistics are important for assessing optimization-based methods. Our pipeline employs a deterministic initialization from the first-frame primitive fit (Section 3.2), yet we recognize that reporting sensitivity strengthens the claims. In the revised manuscript we have augmented the Experiments section and tables with error bars over five random initializations per sequence as well as a sensitivity study on the visibility threshold and joint regularization weights, showing that the performance margins remain consistent. revision: yes

Circularity Check

0 steps flagged

No circularity: optimization framework is self-contained

full rationale

The paper describes a category-agnostic optimization procedure that fits geometric primitives to monocular video, then organizes them into parts via revolute/prismatic joint constraints while jointly solving for segmentation and kinematics. No step reduces a claimed result to its own inputs by definition, renames a fitted quantity as a prediction, or relies on a load-bearing self-citation whose content is itself unverified. The visibility-aware energy and benchmark comparisons are presented as external constraints and evaluations, keeping the derivation independent of the target outputs.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 2 invented entities

Based on abstract only, the claim rests on the domain assumption that primitives plus joint constraints suffice for kinematics recovery; no explicit free parameters or invented entities beyond the new benchmarks are identifiable without full text.

axioms (1)
  • domain assumption Geometric primitives can serve as a proxy representation for articulated parts that enables stable optimization under occlusions and ego-motion
    Invoked as the core alternative to unstable point tracks in the framework description.
invented entities (2)
  • AiP-synth benchmark no independent evidence
    purpose: Synthetic dataset featuring significant camera motion and heavy occlusions for evaluation
    New dataset introduced by authors to demonstrate outperformance.
  • AiP-real benchmark no independent evidence
    purpose: Real dataset featuring significant camera motion and heavy occlusions for evaluation
    New dataset introduced by authors to demonstrate outperformance.

pith-pipeline@v0.9.0 · 5730 in / 1518 out tokens · 84216 ms · 2026-05-20T11:57:18.025503+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

49 extracted references · 49 canonical work pages

  1. [1]

    In: International Conference for Learning Representations (2026)

    Ai, H., Chang, W., Jiao, J., Leonardis, A., Ofek, E.: Articulation in Motion: Prior-free Part Mobility Analysis for Articulated Objects By Dynamic-Static Disentanglement. In: International Conference for Learning Representations (2026)

    work page 2026

  2. [2]

    In: International Conference on Computer Vision Workshops (2025)

    Artykov, A., Boittiaux, C., Lepetit, V .: Articulated Object Understanding from a Single Video Sequence. In: International Conference on Computer Vision Workshops (2025)

    work page 2025

  3. [3]

    IEEE Computer graphics and Applications1(1), 11–23 (1981)

    Barr, A.H.: Superquadrics and angle-preserving transformations. IEEE Computer graphics and Applications1(1), 11–23 (1981)

    work page 1981

  4. [4]

    In: British Machine Vision Conference (2025)

    Chao, J.J., Jiang, Q., Isler, V .: Part Segmentation and Motion Estimation for Articulated Objects with Dynamic 3D Gaussians. In: British Machine Vision Conference (2025)

    work page 2025

  5. [5]

    arXiv preprint arXiv:2603.22102 (2026)

    Dai, H., Fan, H., Zhang, H., Wu, D., Zhang, J., Dong, H.: Freeartgs: Articulated gaussian splatting under free-moving scenario. arXiv preprint arXiv:2603.22102 (2026)

    work page arXiv 2026

  6. [6]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Deng, B., Genova, K., Yazdani, S., Bouaziz, S., Hinton, G., Tagliasacchi, A.: Cvxnet: Learnable convex decomposition. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 31–44 (2020)

    work page 2020

  7. [7]

    In: Proceedings of the IEEE/CVF International Conference on Computer Vision

    Fedele, E., Sun, B., Guibas, L., Pollefeys, M., Engelmann, F.: Superdec: 3d scene decomposition with superquadrics primitives. In: Proceedings of the IEEE/CVF International Conference on Computer Vision. pp. 24625–24635 (2025)

    work page 2025

  8. [8]

    arXiv preprint arXiv:2512.09201 (2025)

    Ganeshan, A., Gadelha, M., Groueix, T., Chen, Z., Chaudhuri, S., Kim, V ., Yifan, W., Ritchie, D.: Residual primitive fitting of 3d shapes with superfrusta. arXiv preprint arXiv:2512.09201 (2025)

    work page arXiv 2025

  9. [9]

    In: Proceedings of the IEEE conference on computer vision and pattern recognition

    Groueix, T., Fisher, M., Kim, V .G., Russell, B.C., Aubry, M.: A papier-mâché approach to learning 3d surface generation. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 216–224 (2018)

    work page 2018

  10. [10]

    In: arXiv Preprint (2025)

    Huang, J., Zhou, Q., Rabeti, H., Korovko, A., Ling, H., Ren, X., Shen, T., Gao, J., Slepichev, D., Lin, C.H., Ren, J., Xie, K., Biswas, J., Leal-Taixe, L., Fidler, S.: ViPE: Video Pose Engine for 3D Geometric Perception. In: arXiv Preprint (2025)

    work page 2025

  11. [11]

    In: Conference on Computer Vision and Pattern Recognition (2022)

    Jiang, Z., Hsu, C.C., Zhu, Y .: Ditto: Building Digital Twins of Articulated Objects from Interaction. In: Conference on Computer Vision and Pattern Recognition (2022)

    work page 2022

  12. [12]

    In: ACM SIGGRAPH (2023)

    Kerbl, B., Kopanas, G., Leimkühler, T., Drettakis, G.: 3D Gaussian Splatting for Real-Time Radiance Field Rendering. In: ACM SIGGRAPH (2023)

    work page 2023

  13. [13]

    In: CORL (2024)

    Kerr, J., Kim, C.M., Wu, M., Yi, B., Wang, Q., Goldberg, K., Kanazawa, A.: Robot See Robot Do: Imitating Articulated Object Manipulation with Monocular 4D Reconstruction. In: CORL (2024)

    work page 2024

  14. [14]

    In: International Confer- ence on Machine Learning (2015)

    Kingma, D.P., Ba, J.: Adam: A Method for Stochastic Optimization. In: International Confer- ence on Machine Learning (2015)

    work page 2015

  15. [15]

    ACM Transactions on Graphics39(6) (2020)

    Laine, S., Hellsten, J., Karras, T., Seol, Y ., Lehtinen, J., Aila, T.: Modular primitives for high-performance differentiable rendering. ACM Transactions on Graphics39(6) (2020)

    work page 2020

  16. [16]

    In: International Conference for Learning Representations (2024) 10

    Le, L., Xie, J., Liang, W., Wang, H.J., Yang, Y ., Ma, Y .J., Vedder, K., Krishna, A., Jayaraman, D., Eaton, E.: Articulate-Anything: Automatic Modeling of Articulated Objects via a Vision- Language Foundation Model. In: International Conference for Learning Representations (2024) 10

    work page 2024

  17. [17]

    In: Conference on Computer Vision and Pattern Recognition (2020)

    Li, X., Wang, H., Yi, L., Guibas, L.J., Abbott, A.L., Song, S.: Category-Level Articulated Object Pose Estimation. In: Conference on Computer Vision and Pattern Recognition (2020)

    work page 2020

  18. [18]

    In: International Conference on Computer Vision (2023)

    Liu, J., Mahdavi-Amiri, A., Savva, M.: Paris: Part-Level Reconstruction and Motion Analysis for Articulated Objects. In: International Conference on Computer Vision (2023)

    work page 2023

  19. [19]

    In: Conference on Computer Vision and Pattern Recognition (2023)

    Liu, S., Gupta, S., Wang, S.: Building Rearticulable Models for Arbitrary 3D Objects from 4D Point Clouds. In: Conference on Computer Vision and Pattern Recognition (2023)

    work page 2023

  20. [20]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Liu, W., Wu, Y ., Ruan, S., Chirikjian, G.S.: Robust and accurate superquadric recovery: A probabilistic approach. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 2676–2685 (2022)

    work page 2022

  21. [21]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Liu, W., Wu, Y ., Ruan, S., Chirikjian, G.S.: Marching-primitives: Shape abstraction from signed distance function. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 8771–8780 (2023)

    work page 2023

  22. [22]

    In: ECCV (2020)

    Mildenhall, B., Srinivasan, P.P., Tancik, M., Barron, J.T., Ramamoorthi, R., Ng, R.: Nerf: Representing scenes as neural radiance fields for view synthesis. In: ECCV (2020)

    work page 2020

  23. [23]

    In: Conference on Computer Vision and Pattern Recognition (2019)

    Mo, K., Zhu, S., Chang, A.X., Yi, L., Tripathi, S., Guibas, L.J., Su, H.: PartNet: A Large- Scale Benchmark for Fine-Grained and Hierarchical Part-Level 3D Object Understanding. In: Conference on Computer Vision and Pattern Recognition (2019)

    work page 2019

  24. [24]

    Advances in Neural Information Processing Systems36, 5791–5807 (2023)

    Monnier, T., Austin, J., Kanazawa, A., Efros, A., Aubry, M.: Differentiable blocks world: Qual- itative 3d decomposition by rendering primitives. Advances in Neural Information Processing Systems36, 5791–5807 (2023)

    work page 2023

  25. [25]

    In: International Conference on Computer Vision (2021)

    Mu, J., Qiu, W., Kortylewski, A., Yuille, A., Vasconcelos, N., Wang, X.: A-SDF: Learning Disentangled Signed Distance Functions for Articulated Shape Representation. In: International Conference on Computer Vision (2021)

    work page 2021

  26. [26]

    In: Conference on Computer Vision and Pattern Recognition (2022)

    Noguchi, A., Iqbal, U., Tremblay, J., Harada, T., Gallo, O.: Watch It Move: Unsupervised Discovery of 3D Joints for Re-Posing of Articulated Objects. In: Conference on Computer Vision and Pattern Recognition (2022)

    work page 2022

  27. [27]

    In: International Conference on Robotics and Automa- tion (2023)

    Ota, K., Tung, H.Y ., Smith, K.A., Cherian, A., Marks, T.K., Sullivan, A., Kanezaki, A., Tenen- baum, J.B.: H-SAUR: Hypothesize, Simulate, Act, Update, and Repeat for Understanding Object Articulations from Interactions. In: International Conference on Robotics and Automa- tion (2023)

    work page 2023

  28. [28]

    In: Conference on Computer Vision and Pattern Recognition (2019)

    Park, J.J., Florence, P., Straub, J., Newcombe, R.A., Lovegrove, S.J.: DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation. In: Conference on Computer Vision and Pattern Recognition (2019)

    work page 2019

  29. [29]

    In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    Paschalidou, D., Katharopoulos, A., Geiger, A., Fidler, S.: Neural parts: Learning expressive 3d shape abstractions with invertible neural networks. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. pp. 3204–3215 (2021)

    work page 2021

  30. [30]

    In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Paschalidou, D., Ulusoy, A.O., Geiger, A.: Superquadrics revisited: Learning 3d shape parsing beyond cuboids. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 10344–10353 (2019)

    work page 2019

  31. [31]

    In: International Conference on 3D Vision (2026)

    Peng, W., Lv, J., Lu, C., Savva, M.: iTACO: Interactable Digital Twins of Articulated Objects from Casually Captured RGBD Videos. In: International Conference on 3D Vision (2026)

    work page 2026

  32. [32]

    Pilu, M., Fisher, R.B.: Equal-distance sampling of superellipse models (1995)

    work page 1995

  33. [33]

    In: arXiv Preprint (2024)

    Ravi, N., Gabeur, V ., Hu, Y .T., Hu, R., Ryali, C., Ma, T., Khedr, H., Rädle, R., Rolland, C., Gustafson, L., Mintun, E., Pan, J., Alwala, K.V ., Carion, N., Wu, C.Y ., Girshick, R., Dollár, P., Feichtenhofer, C.: SAM 2: Segment Anything in Images and Videos. In: arXiv Preprint (2024)

    work page 2024

  34. [34]

    In: American Association for Artificial Intelligence Conference (2022) 11

    Shi, Y ., Cao, X., Lu, F., Zhou, B.: P3-Net: Part Mobility Parsing from Point Cloud Sequences via Learning Explicit Point Correspondence. In: American Association for Artificial Intelligence Conference (2022) 11

    work page 2022

  35. [35]

    In: Computer Graphics Forum (2021)

    Shi, Y ., Cao, X., Zhou, B.: Self-Supervised Learning of Part Mobility from Point Cloud Sequence. In: Computer Graphics Forum (2021)

    work page 2021

  36. [36]

    In: Conference on Computer Vision and Pattern Recognition (2024)

    Song, C., Wei, J., Foo, C.S., Lin, G., Liu, F.: REACTO: Reconstructing Articulated Objects from a Single Video. In: Conference on Computer Vision and Pattern Recognition (2024)

    work page 2024

  37. [37]

    In: Proceedings of the IEEE conference on computer vision and pattern recognition

    Tulsiani, S., Su, H., Guibas, L.J., Efros, A.A., Malik, J.: Learning shape abstractions by assembling volumetric primitives. In: Proceedings of the IEEE conference on computer vision and pattern recognition. pp. 2635–2643 (2017)

    work page 2017

  38. [38]

    IEEE Transactions on Pattern Analysis and Machine Intelligence41(1), 220–233 (2019)

    Vaskevicius, N., Birk, A.: Revisiting superquadric fitting: A numerically stable formulation. IEEE Transactions on Pattern Analysis and Machine Intelligence41(1), 220–233 (2019)

    work page 2019

  39. [39]

    In: Conference on Computer Vision and Pattern Recognition (2019)

    Wang, H., Sridhar, S., Huang, J., Valentin, J., Song, S., Guibas, L.J.: Normalized Object Coordinate Space for Category-Level 6D Object Pose and Size Estimation. In: Conference on Computer Vision and Pattern Recognition (2019)

    work page 2019

  40. [40]

    In: International Conference on Computer Vision (2021)

    Weng, Y ., Wang, H., Zhou, Q., Qin, Y ., Duan, Y ., Fan, Q., Chen, B., Su, H., Guibas, L.J.: CAPTRA: CAtegory-level Pose Tracking for Rigid and Articulated Objects from Point Clouds. In: International Conference on Computer Vision (2021)

    work page 2021

  41. [41]

    In: Conference on Computer Vision and Pattern Recognition (2024)

    Weng, Y ., Wen, B., Tremblay, J., Blukis, V ., Fox, D., Guibas, L., Birchfield, S.: Neural Implicit Representation for Building Digital Twins of Unknown Articulated Objects. In: Conference on Computer Vision and Pattern Recognition (2024)

    work page 2024

  42. [42]

    In: CoRL (2025)

    Werby, A., Buechner, M., Roefer, A., Huang, C., Burgard, W., Valada, A.: Articulated Object Estimation in the Wild. In: CoRL (2025)

    work page 2025

  43. [43]

    In: Advances in Neural Information Processing Systems (2025)

    Wu, D., Liu, L., Linli, Z., Huang, A., Song, L., Yu, Q., Wu, Q., Lu, C.: Reartgs: Reconstructing and Generating Articulated Objects via 3D Gaussian Splatting with Geometric and Motion Constraints. In: Advances in Neural Information Processing Systems (2025)

    work page 2025

  44. [44]

    In: Conference on Computer Vision and Pattern Recognition (2020)

    Xiang, F., Qin, Y ., Mo, K., Xia, Y ., Zhu, H., Liu, F., Liu, M., Jiang, H., Yuan, Y ., Wang, H., Yi, L., Chang, A.X., Guibas, L.J., Su, H.: SAPIEN: A SimulAted Part-Based Interactive ENvironment. In: Conference on Computer Vision and Pattern Recognition (2020)

    work page 2020

  45. [45]

    ACM Transactions On Graphics (TOG)40(4), 1–11 (2021)

    Yang, K., Chen, X.: Unsupervised learning for cuboid shape abstraction via joint segmentation from point clouds. ACM Transactions On Graphics (TOG)40(4), 1–11 (2021)

    work page 2021

  46. [46]

    In: Advances in Neural Information Processing Systems (2023)

    Zhang, Y ., Edstedt, J., Wandt, B., Forssén, P.E., Magnusson, M., Felsberg, M.: GMSF: Global Matching Scene Flow. In: Advances in Neural Information Processing Systems (2023)

    work page 2023

  47. [47]

    In: Advances in Neural Information Processing Systems (2023)

    Zhong, J.X., Cheng, T.Y ., He, Y ., Lu, K., Zhou, K., Markham, A., Trigoni, N.: Multi-Body SE(3) Equivariance for Unsupervised Rigid Segmentation and Motion Estimation. In: Advances in Neural Information Processing Systems (2023)

    work page 2023

  48. [48]

    Limitations

    Zhou, Y ., Barnes, C., Lu, J., Yang, J., Li, H.: On the continuity of rotation representations in neural networks. In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. pp. 5745–5753 (2019) 12 Appendix Articulation in Prime: Primitive-Based Articulated Object Understanding from a Single Casual Video A Superquadrics Superqu...

    work page 2019

  49. [49]

    Guidelines: • The answer [N/A] means that the paper does not involve crowdsourcing nor research with human subjects

    Institutional review board (IRB) approvals or equivalent for research with human subjects Question: Does the paper describe potential risks incurred by study participants, whether such risks were disclosed to the subjects, and whether Institutional Review Board (IRB) approvals (or an equivalent approval/review based on the requirements of your country or ...

    work page