arxiv: 2604.20800 · v1 · submitted 2026-04-22 · 💻 cs.CV · cs.LG

Recognition: unknown

LEXIS: LatEnt ProXimal Interaction Signatures for 3D HOI from an Image

Dimitrije Anti\'c , Alvaro Budria , George Paschalidis , Sai Kumar Dwivedi , Dimitrios Tzionas

Authors on Pith no claims yet

Pith reviewed 2026-05-10 00:42 UTC · model grok-4.3

classification 💻 cs.CV cs.LG

keywords 3D human-object interactionsingle-image reconstructionproximal interaction signaturesdense proximity fieldsVQ-VAEdiffusion modelsInterFieldsphysical plausibility

0 comments

The pith

A learned discrete manifold of proximal interaction signatures enables single-image inference of 3D human-object meshes together with their dense continuous proximity fields.

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

Reconstructing 3D human-object interactions from a single RGB image requires modeling how bodies and objects approach each other across entire surfaces rather than at isolated contact points. Prior methods rely on sparse binary contact labels that miss the gradual closeness typical of real scenes. The paper introduces InterFields as a dense, continuous representation of proximity between body and object surfaces. It captures the structure of typical interactions by training a VQ-VAE on 3D data to obtain LEXIS, a compact discrete manifold of interaction signatures. A diffusion model conditioned on this manifold then generates both meshes and InterFields directly from the image, guiding refinement to produce physically plausible outputs without separate optimization.

Core claim

The paper claims that interaction patterns are structured by action type and object geometry, allowing a vector-quantized variational autoencoder to learn a useful discrete manifold called LEXIS of proximal interaction signatures. Conditioned on LEXIS, the LEXIS-Flow diffusion framework reconstructs human and object meshes along with dense InterFields from one image; the inferred fields then steer a guided refinement step that enforces realistic proximity without post-hoc optimization.

What carries the argument

LEXIS, the discrete manifold of proximal interaction signatures obtained via VQ-VAE; it supplies structured conditioning that lets the diffusion model infer dense InterFields from image input.

If this is right

Mesh accuracy, contact precision, and proximity estimation all improve over existing single-image baselines on the Open3DHOI and BEHAVE datasets.
Generalization to unseen actions and objects increases while the generated scenes are rated more realistic by human observers.
Physically plausible results emerge directly from the generation process because InterFields guide refinement without any separate optimization stage.
The approach moves 3D scene understanding closer to capturing continuous physical relationships between people and objects.

Where Pith is reading between the lines

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

The signatures may implicitly encode action semantics from geometry alone, opening the possibility of action recognition as a byproduct of reconstruction.
The same manifold-learning strategy could be applied to related ill-posed tasks such as hand-object interaction or multi-person contact.
In robotics or AR settings the continuous proximity fields could supply richer distance information for planning grasps or avoiding collisions than binary contacts allow.

Load-bearing premise

Interaction patterns are characteristically structured by the action and object geometry in a manner that permits a VQ-VAE to extract a useful discrete manifold of signatures.

What would settle it

Training an ablated version of LEXIS-Flow that directly regresses InterFields from images without the VQ-VAE manifold and showing equal or superior reconstruction, contact, and proximity metrics on Open3DHOI and BEHAVE would undermine the claimed necessity of the learned signatures.

Figures

Figures reproduced from arXiv: 2604.20800 by Alvaro Budria, Dimitrije Anti\'c, Dimitrios Tzionas, George Paschalidis, Sai Kumar Dwivedi.

**Figure 1.** Figure 1: We present LEXIS-Flow, a framework for 3D Human-Object Interaction (HOI ) reconstruction from single images. We go beyond sparse, binary contact by learning LEXIS, a latent manifold of dense, continuous interaction signatures. LEXIS-guided sampling enables recovering physically-plausible HOI without post-hoc optimization. Abstract. Reconstructing 3D Human–Object Interaction from an RGB image is essential f… view at source ↗

**Figure 2.** Figure 2: Toy example. Contact between the body and bench, and between hands and dumbbells, remains fixed. Instead, distances from each body point to the objects, v.v., form a rich, geometry- and proximity-aware interaction signature. Our actions are inherently defined and physically constrained bythe objects around us. Accurately recovering 3D human–object interaction (HOI) from a single RGB image is essential for… view at source ↗

**Figure 3.** Figure 3: LEXIS (Sec. 3.1). To estimate 3D HOI from images we need a prior model of interactions. We train LEXIS-Net, a VQ-VAE that learns a compact dictionary of proximal interaction signatures, termed LEXIS. The encoder Eϕ maps 3D body pose to continuous latents Zˆ, quantized via a learned codebook C into discrete tokens Z, and decoded via Dψ into 3D body pose and body/object InterFields (shown color-coded). Appro… view at source ↗

**Figure 4.** Figure 4: TriPlane-based InterFields. A 3D surface point p (see yellow point on human/object surface) is orthogonally projected onto the three feature planes (see red, blue, and green points) to sample features, which are aggregated and passed to a small MLP to infer the InterField value for point p. \triplaneHum ,\, \triplaneObj = \lexisDecoder (\lexisTokens ,\, \objfeat ), \label {eq:triplane_decode} \vspace {+… view at source ↗

**Figure 5.** Figure 5: LEXIS-Flow (Sec. 3.2). We develop a dual-stream Flow-Matching model that takes as input a single image, and estimates 3D body and object meshes in interaction along with LEXIS-based InterField proximal relationships (Sec. 3.1; shown with heatmap color-coding on the 3D meshes). Guiding sampling via LEXIS-based InterFields refines estimates to improves the physical plausibility of 3D interaction. 3.2 LEXIS-… view at source ↗

**Figure 6.** Figure 6: LEXIS-Flow vs SotA. Existing methods often fail to capture tight physical coupling, yielding floating objects (HOI-Gaussian, InteractVLM) or penetrations (HDM). LEXIS-Flow tackles this via dense InterFields for proximity-aware estimation. Guided Refinement (Tab. 1): With identical initialization (row D–G) LEXIS-Flow∗ (row G) achieves the best results across all four metrics, outperforming both InteractVL… view at source ↗

**Figure 7.** Figure 7: Qualitative results. LEXIS-Flow∗ recovers physically-plausible interactions from diverse in-the-wild images. By leveraging dense InterField signals, our model produces accurate spatial configurations and realistic articulations even under occlusion. 4.5 Ablation Study Variants CDhum ↓ CDobj ↓ Architect. (Unguided) GaussFlow Baseline 13.45 57.25 FlowObject Baseline N/A 65.05 LEXIS-Flow Baseline 9.68 48.01 … view at source ↗

read the original abstract

Reconstructing 3D Human-Object Interaction from an RGB image is essential for perceptive systems. Yet, this remains challenging as it requires capturing the subtle physical coupling between the body and objects. While current methods rely on sparse, binary contact cues, these fail to model the continuous proximity and dense spatial relationships that characterize natural interactions. We address this limitation via InterFields, a representation that encodes dense, continuous proximity across the entire body and object surfaces. However, inferring these fields from single images is inherently ill-posed. To tackle this, our intuition is that interaction patterns are characteristically structured by the action and object geometry. We capture this structure in LEXIS, a novel discrete manifold of interaction signatures learned via a VQ-VAE. We then develop LEXIS-Flow, a diffusion framework that leverages LEXIS signatures to estimate human and object meshes alongside their InterFields. Notably, these InterFields help in a guided refinement that ensures physically-plausible, proximity-aware reconstructions without requiring post-hoc optimization. Evaluation on Open3DHOI and BEHAVE shows that LEXIS-Flow significantly outperforms existing SotA baselines in reconstruction, contact, and proximity quality. Our approach not only improves generalization but also yields reconstructions perceived as more realistic, moving us closer to holistic 3D scene understanding. Code & models will be public at https://anticdimi.github.io/lexis.

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

LEXIS learns a discrete manifold of interaction signatures via VQ-VAE to enable diffusion-based reconstruction of 3D HOI with continuous proximity, offering a clear advance over binary contact methods though the necessity of the discretization step remains to be fully tested.

read the letter

The punchline is that this paper takes the problem of 3D human-object interaction reconstruction from images and tries to handle continuous proximity instead of just binary contacts by learning a discrete set of interaction signatures. They define InterFields to capture dense proximity across surfaces and then train a VQ-VAE to turn those into LEXIS, a codebook of signatures. LEXIS-Flow is the diffusion model that uses this to generate the human and object meshes plus the fields from a single image. They add a guided refinement step to keep things physically plausible. The paper does well at spotting the gap in current methods that stick to sparse contacts and at proposing a representation that directly models the continuous nature of real interactions. Applying VQ-VAE and diffusion here is not entirely new in CV but the combination for this task and the claim of no post-hoc optimization needed is a plus. The results on Open3DHOI and BEHAVE suggest better performance on reconstruction, contact, and proximity, with more realistic outputs and some generalization improvement. Where it could be softer is in validating that the discrete manifold is actually doing the heavy lifting. The assumption that actions and object shapes create characteristic patterns that a VQ-VAE can usefully quantize might work for the tested cases but could break on more varied or novel interactions. Without seeing the full methods and ablations, it's unclear how much the improvements come from the LEXIS component versus the diffusion model or the refinement. The ill-posed nature means the learned prior is critical, so any overfitting to the training distributions on those two datasets would limit broader use. Also, diffusion models often require careful tuning, and the free parameters listed like codebook size could influence outcomes more than acknowledged. This paper is for computer vision researchers focused on 3D reconstruction of interactions and anyone building systems that need accurate physical scene understanding, like in robotics. A reader looking for new ways to represent proximity in HOI would get something concrete to try or extend. It deserves a serious referee because the core idea is coherent and the evaluation uses public data with clear metrics, even if more experiments would strengthen it. I recommend sending this to peer review. The work is solid enough on its own terms to warrant discussion, though reviewers will likely push for ablations on the VQ-VAE and checks on generalization.

Referee Report

0 major / 2 minor

Summary. The paper introduces InterFields as a dense, continuous proximity representation for 3D human-object interactions (HOI) from a single RGB image, contrasting with prior sparse binary contact cues. It proposes LEXIS, a discrete manifold of interaction signatures learned via VQ-VAE that captures structure from actions and object geometry, and LEXIS-Flow, a diffusion framework that infers human/object meshes together with InterFields; these fields then guide refinement for physically plausible outputs without post-hoc optimization. The central empirical claim is that LEXIS-Flow significantly outperforms existing SOTA baselines on Open3DHOI and BEHAVE in reconstruction, contact, and proximity quality while improving generalization and perceived realism.

Significance. If the performance claims hold, the work advances 3D HOI reconstruction by replacing sparse cues with continuous proximity fields and by using a learned discrete manifold to regularize an otherwise ill-posed inverse problem. The diffusion-based inference with guided refinement is a practical contribution, and the explicit commitment to release code and models supports reproducibility. The approach could influence downstream tasks in scene understanding that require physically grounded 3D interactions.

minor comments (2)

The abstract states that InterFields 'help in a guided refinement' but does not indicate whether this step is part of the diffusion sampling loop or a separate post-processing stage; a brief clarification would aid readers.
Evaluation is reported on Open3DHOI and BEHAVE, yet the abstract does not name the quantitative metrics (e.g., Chamfer distance, contact IoU, proximity error) used to support the 'significantly outperforms' claim; adding these would strengthen the summary.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the careful review and for recognizing the potential significance of InterFields as a continuous proximity representation and LEXIS as a learned discrete manifold for regularizing 3D HOI reconstruction. We appreciate the positive assessment of the diffusion framework with guided refinement and the commitment to code release. The 'uncertain' recommendation appears to stem from the need for further validation of the performance claims; below we provide point-by-point clarification on the major aspects raised in the summary, while noting that no specific technical criticisms were detailed in the major comments section.

Circularity Check

0 steps flagged

No significant circularity; empirical ML pipeline on public data

full rationale

The paper introduces InterFields as a dense proximity representation and LEXIS as a VQ-VAE-learned discrete manifold of signatures, then applies a diffusion model (LEXIS-Flow) to infer meshes and fields from images. All components are trained end-to-end on external datasets (Open3DHOI, BEHAVE) with quantitative evaluation against baselines. No equations or steps reduce by construction to fitted inputs, self-definitions, or self-citation chains; the manifold is data-driven rather than presupposed, and performance claims rest on held-out empirical metrics rather than analytic closure. The framework is self-contained against external benchmarks and falsifiable via standard ML protocols.

Axiom & Free-Parameter Ledger

2 free parameters · 1 axioms · 2 invented entities

The central claim relies on these new invented representations and the assumption about structured patterns. No independent evidence for the entities beyond the model's performance is provided in the abstract. Free parameters are implicit in the neural network architectures.

free parameters (2)

VQ-VAE codebook size and training hyperparameters
Typical in VQ-VAE for learning LEXIS signatures, not specified in abstract.
Diffusion model architecture parameters
Many parameters in LEXIS-Flow not detailed in abstract.

axioms (1)

domain assumption Interaction patterns are characteristically structured by the action and object geometry.
This intuition justifies the use of LEXIS signatures to address the ill-posed inference problem.

invented entities (2)

InterFields no independent evidence
purpose: Representation encoding dense continuous proximity across body and object surfaces
Newly proposed to capture more than binary contacts.
LEXIS no independent evidence
purpose: Discrete manifold of interaction signatures learned via VQ-VAE
New discrete manifold proposed as structured prior.

pith-pipeline@v0.9.0 · 5570 in / 1592 out tokens · 41164 ms · 2026-05-10T00:42:19.673336+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

95 extracted references · 3 canonical work pages · 2 internal anchors

[1]

In: International Conference on Computer Vision (ICCV)

Antić, D., Paschalidis, G., Tripathi, S., Gevers, T., Dwivedi, S.K., Tzionas, D.: SDFit: 3D object pose and shape by fitting a morphable SDF to a single image. In: International Conference on Computer Vision (ICCV). pp. 9616–9626 (2025) 4

2025
[2]

In: International Conference on Machine Learning (ICML)

Bao, F., Nie, S., Xue, K., Li, C., Pu, S., Wang, Y., Yue, G., Cao, Y., Su, H., Zhu, J.: One transformer fits all distributions in multi-modal diffusion at scale. In: International Conference on Machine Learning (ICML). pp. 1692–1717 (2023) 8

2023
[3]

Transactions on Robotics (T-RO)21(1), 47–57 (2005) 4

Bernardin, K., Ogawara, K., Ikeuchi, K., Dillmann, R.: A sensor fusion approach for recognizing continuous human grasping sequences using hidden markov models. Transactions on Robotics (T-RO)21(1), 47–57 (2005) 4

2005
[4]

In: Computer Vision and Pattern Recognition (CVPR)

Bhatnagar, B.L., Xie, X., Petrov, I.A., Sminchisescu, C., Theobalt, C., Pons-Moll, G.: BEHAVE: Dataset and method for tracking human object interactions. In: Computer Vision and Pattern Recognition (CVPR). pp. 15914–15925 (2022) 3, 4, 11, 13

2022
[5]

In: European Conference on Computer Vision (ECCV)

Bogo, F., Kanazawa, A., Lassner, C., Gehler, P., Romero, J., Black, M.J.: Keep it SMPL: Automatic estimation of 3D human pose and shape from a single image. In: European Conference on Computer Vision (ECCV). vol. 9909, pp. 561–578 (2016) 4

2016
[6]

In: European Conference on Computer Vision (ECCV)

Brahmbhatt, S., Tang, C., Twigg, C.D., Kemp, C.C., Hays, J.: ContactPose: A dataset of grasps with object contact and hand pose. In: European Conference on Computer Vision (ECCV). vol. 12358, pp. 361–378 (2020) 2

2020
[7]

In: Computer Vision and Pat- tern Recognition (CVPR)

Chan, E.R., Lin, C.Z., Chan, M.A., Nagano, K., Pan, B., Mello, S.D., Gallo, O., Guibas, L.J., Tremblay, J., Khamis, S., Karras, T., Wetzstein, G.: Efficient geometry-aware 3D generative adversarial networks. In: Computer Vision and Pat- tern Recognition (CVPR). pp. 16102–16112 (2022) 7

2022
[8]

Chen, Y., Dwivedi, S.K., Black, M.J., Tzionas, D.: Detecting human-object contact inimages.In:ComputerVisionandPatternRecognition(CVPR).pp.17100–17110 (2023) 4

2023
[9]

In: Computer Vision and Pattern Recognition (CVPR)

Chen, Z., Zhang, H.: Learning implicit fields for generative shape modeling. In: Computer Vision and Pattern Recognition (CVPR). pp. 5939–5948 (2019) 4

2019
[10]

In: Computer Vision and Pattern Recognition (CVPR)

Cheng, Y.C., Lee, H.Y., Tuyakov, S., Schwing, A., Gui, L.: SDFusion: Multimodal 3D shape completion, reconstruction, and generation. In: Computer Vision and Pattern Recognition (CVPR). pp. 4456–4465 (2023) 4

2023
[11]

In: European Conference on Computer Vision (ECCV)

Choy, C.B., Xu, D., Gwak, J., Chen, K., Savarese, S.: 3D-R2N2: A unified approach for single and multi-view 3D object reconstruction. In: European Conference on Computer Vision (ECCV). vol. 9912, pp. 628–644 (2016) 4

2016
[12]

Corona, E., Pons-Moll, G., Alenyà, G., Moreno-Noguer, F.: Learned vertex descent: Anewdirectionfor3Dhumanmodelfitting.In:EuropeanConferenceonComputer Vision (ECCV). pp. 146–165 (2022) 4

2022
[13]

In: Computer Vision and Pattern Recognition (CVPR)

Cseke, A., Tripathi, S., Dwivedi, S.K., Lakshmipathy, A., Chatterjee, A., Black, M.J., Tzionas, D.: PICO: Reconstructing 3D people in contact with objects. In: Computer Vision and Pattern Recognition (CVPR). pp. 1783–1794 (2025) 4, 5

2025
[14]

In: Computer Vision and Pattern Recognition (CVPR)

Deitke, M., Schwenk, D., Salvador, J., Weihs, L., Michel, O., VanderBilt, E., Schmidt, L., Ehsani, K., Kembhavi, A., Farhadi, A.: Objaverse: A universe of annotated 3D objects. In: Computer Vision and Pattern Recognition (CVPR). pp. 13142–13153 (2023) 5

2023
[15]

In: Computer Vision and Pattern Recognition (CVPR)

Diller, C., Dai, A.: CG-HOI: Contact-guided 3D human-object interaction gener- ation. In: Computer Vision and Pattern Recognition (CVPR). pp. 19888–19901 (2024) 4 LEXIS 17

2024
[16]

In: Com- puter Vision and Pattern Recognition (CVPR)

Dwivedi, S.K., Antić, D., Tripathi, S., Taheri, O., Schmid, C., Black, M.J., Tzionas, D.: InteractVLM: 3D interaction reasoning from 2D foundational models. In: Com- puter Vision and Pattern Recognition (CVPR). pp. 22605–22615 (2025) 2, 3, 4, 5, 11, 12, 13

2025
[17]

In: International Conference on 3D Vision (3DV)

Dwivedi, S.K., Schmid, C., Yi, H., Black, M.J., Tzionas, D.: POCO: 3D pose and shape estimation using confidence. In: International Conference on 3D Vision (3DV). pp. 85–95 (2024) 4

2024
[18]

In: Computer Vision and Pattern Recognition (CVPR)

Dwivedi, S.K., Sun, Y., Patel, P., Feng, Y., Black, M.J.: TokenHMR: Advancing human mesh recovery with a tokenized pose representation. In: Computer Vision and Pattern Recognition (CVPR). pp. 1323–1333 (2024) 3, 6, 7, 10

2024
[19]

In: Computer Vision and Pattern Recognition (CVPR)

Fan, H., Su, H., Guibas, L.: A point set generation network for 3D object re- construction from a single image. In: Computer Vision and Pattern Recognition (CVPR). pp. 2463–2471 (2017) 4

2017
[20]

In: Computer Vision and Pattern Recognition (CVPR)

Fan, Z., Taheri, O., Tzionas, D., Kocabas, M., Kaufmann, M., Black, M.J., Hilliges, O.: ARCTIC: A dataset for dexterous bimanual hand-object manipulation. In: Computer Vision and Pattern Recognition (CVPR). pp. 12943–12954 (2023) 2, 4

2023
[21]

Transactions on Human-Machine Systems (THMS)46(1), 66–77 (2016) 4

Feix, T., Romero, J., Schmiedmayer, H.B., Dollar, A.M., Kragic, D.: The GRASP taxonomy of human grasp types. Transactions on Human-Machine Systems (THMS)46(1), 66–77 (2016) 4

2016
[22]

In: European Conference on Computer Vision (ECCV)

Fiche, G., Leglaive, S., Alameda-Pineda, X., Agudo, A., Moreno-Noguer, F.: VQ- HPS: Human pose and shape estimation in a vector-quantized latent space. In: European Conference on Computer Vision (ECCV). vol. 15110, pp. 471–490 (2024) 3

2024
[23]

In: International Conference on Computer Vision (ICCV)

Gkioxari, G., Malik, J., Johnson, J.: Mesh R-CNN. In: International Conference on Computer Vision (ICCV). pp. 9784–9794 (2019) 4

2019
[24]

In: European Conference on Computer Vision (ECCV)

Goodwin,W.,Vaze,S.,Havoutis,I.,Posner,I.:Zero-shotcategory-levelobjectpose estimation. In: European Conference on Computer Vision (ECCV). vol. 13699, pp. 516–532 (2022) 4

2022
[25]

In: European Conference on Computer Vision (ECCV)

Grady, P., Tang, C., Brahmbhatt, S., Twigg, C.D., Wan, C., Hays, J., Kemp, C.C.: PressureVision: Estimating hand pressure from a single RGB image. In: European Conference on Computer Vision (ECCV). vol. 13666, pp. 328–345 (2022) 4

2022
[26]

In: Computer Vision and Pattern Recognition (CVPR)

Grady, P., Tang, C., Twigg, C.D., Vo, M., Brahmbhatt, S., Kemp, C.C.: Contac- tOpt: Optimizing contact to improve grasps. In: Computer Vision and Pattern Recognition (CVPR). pp. 1471–1481 (2021) 4

2021
[27]

In: Computer Vision and Pattern Recognition (CVPR)

Hampali, S., Rad, M., Oberweger, M., Lepetit, V.: HOnnotate: A method for 3D annotation of hand and object poses. In: Computer Vision and Pattern Recognition (CVPR). pp. 3196–3206 (2020) 4

2020
[28]

In: International Conference on Com- puter Vision (ICCV)

Han, S., Joo, H.: CHORUS: Learning canonicalized 3D human-object spatial rela- tions from unbounded synthesized images. In: International Conference on Com- puter Vision (ICCV). pp. 15789–15800 (2023) 4

2023
[29]

In: International Conference on Computer Vision (ICCV)

Hassan, M., Choutas, V., Tzionas, D., Black, M.J.: Resolving 3D human pose ambiguities with 3D scene constraints. In: International Conference on Computer Vision (ICCV). pp. 2282–2292 (2019) 4

2019
[30]

In: Computer Vision and Pattern Recognition (CVPR)

Hassan, M., Ghosh, P., Tesch, J., Tzionas, D., Black, M.J.: Populating 3D scenes by learning human-scene interaction. In: Computer Vision and Pattern Recognition (CVPR). pp. 14708–14718 (2021) 4

2021
[31]

In: Computer Vision and Pattern Recognition (CVPR)

Huang, C.H.P., Yi, H., Höschle, M., Safroshkin, M., Alexiadis, T., Polikovsky, S., Scharstein, D., Black, M.J.: Capturing and inferring dense full-body human-scene contact. In: Computer Vision and Pattern Recognition (CVPR). pp. 13274–13285 (2022) 4 18 D. Antić et al

2022
[32]

International Journal of Computer Vision (IJCV)132(7), 2551–2566 (2024) 4

Huang, Y., Taheri, O., Black, M.J., Tzionas, D.: InterCap: Joint markerless 3D tracking of humans and objects in interaction from multi-view RGB-D images. International Journal of Computer Vision (IJCV)132(7), 2551–2566 (2024) 4

2024
[33]

In: Computer Vision and Pattern Recognition (CVPR)

Huang, Z., Jampani, V., Thai, A., Li, Y., Stojanov, S., Rehg, J.M.: ShapeClipper: Scalable 3D shape learning from single-view images via geometric and CLIP-based consistency. In: Computer Vision and Pattern Recognition (CVPR). pp. 12912– 12922 (2023) 4

2023
[34]

In: International Conference on Computer Vision (ICCV)

Jiang, H., Liu, S., Wang, J., Wang, X.: Hand-object contact consistency reasoning for human grasps generation. In: International Conference on Computer Vision (ICCV). pp. 11087–11096 (2021) 4

2021
[35]

In: Computer Vision and Pattern Recognition (CVPR)

Joo, H., Simon, T., Sheikh, Y.: Total capture: A 3D deformation model for tracking faces, hands, and bodies. In: Computer Vision and Pattern Recognition (CVPR). pp. 8320–8329 (2018) 4

2018
[36]

American Journal of Occupational Therapy34(7), 437–445 (1980) 4

Kamakura, N., Matsuo, M., Ishii, H., Mitsuboshi, F., Miura, Y.: Patterns of static prehension in normal hands. American Journal of Occupational Therapy34(7), 437–445 (1980) 4

1980
[37]

In: Computer Vision and Pattern Recognition (CVPR)

Kanazawa, A., Black, M.J., Jacobs, D.W., Malik, J.: End-to-end recovery of human shape and pose. In: Computer Vision and Pattern Recognition (CVPR). pp. 7122– 7131 (2018) 4

2018
[38]

In: European Conference on Computer Vision (ECCV)

Kim, H., Han, S., Kwon, P., Joo, H.: Beyond the contact: Discovering comprehen- sive affordance for 3D objects from pre-trained 2D diffusion models. In: European Conference on Computer Vision (ECCV). pp. 400–419 (2024) 4

2024
[39]

In: International Conference on Computer Vision (ICCV)

Kirillov, A., Mintun, E., Ravi, N., Mao, H., Rolland, C., Gustafson, L., Xiao, T., Whitehead, S., Berg, A.C., Lo, W.Y., Dollar, P., Girshick, R.: Segment anything. In: International Conference on Computer Vision (ICCV). pp. 4015–4026 (2023) 9, 10

2023
[40]

In: Computer Vision and Pattern Recognition (CVPR)

Kocabas, M., Athanasiou, N., Black, M.: VIBE: Video inference for human body pose and shape estimation. In: Computer Vision and Pattern Recognition (CVPR). pp. 5252–5262 (2020) 4

2020
[41]

Transac- tions on Graphics (TOG)42(6), 197:1–197:11 (2023) 11

Li, J., Wu, J., Liu, C.K.: Object motion guided human motion synthesis. Transac- tions on Graphics (TOG)42(6), 197:1–197:11 (2023) 11

2023
[42]

In: International Conference on Computer Vision (ICCV)

Lin, K., Wang, L., Liu, Z.: Mesh graphormer. In: International Conference on Computer Vision (ICCV). pp. 12919–12928 (2021) 4

2021
[43]

In: International Conference on Learning Representations (ICLR) (2023) 8

Lipman, Y., Chen, R.T.Q., Ben-Hamu, H., Nickel, M., Le, M.: Flow matching for generative modeling. In: International Conference on Learning Representations (ICLR) (2023) 8

2023
[44]

In: Conference on Neural Information Processing Systems (NeurIPS) (2023) 4, 5

Liu, M., Shi, R., Kuang, K., Zhu, Y., Li, X., Han, S., Cai, H., Porikli, F., Su, H.: OpenShape: Scaling up 3D shape representation towards open-world understand- ing. In: Conference on Neural Information Processing Systems (NeurIPS) (2023) 4, 5

2023
[45]

In: Conference on Neural Information Processing Systems (NeurIPS) (2024) 4

Liu, M., Xu, C., Jin, H., Chen, L., Varma T, M., Xu, Z., Su, H.: One-2-3-45: Any single image to 3D mesh in 45 seconds without per-shape optimization. In: Conference on Neural Information Processing Systems (NeurIPS) (2024) 4

2024
[46]

In: International Conference on Computer Vision (ICCV)

Liu, R., Wu, R., Hoorick, B.V., Tokmakov, P., Zakharov, S., Vondrick, C.: Zero-1- to-3: Zero-shot one image to 3D object. In: International Conference on Computer Vision (ICCV). pp. 9264–9275 (2023) 4

2023
[47]

In: International Conference on Computer Vision (ICCV)

Liu, S., Zhou, Y., Yang, J., Gupta, S., Wang, S.: ContactGen: Generative contact modeling for grasp generation. In: International Conference on Computer Vision (ICCV). pp. 20552–20563 (2023) 2, 4 LEXIS 19

2023
[48]

In: International Conference on Computer Vision (ICCV)

Liu, Z., Zhou, D., Lu, F., Fang, J., Zhang, L.: AutoShape: Real-time shape-aware monocular 3D object detection. In: International Conference on Computer Vision (ICCV). pp. 15621–15630 (2021) 4

2021
[49]

Transactions on Graphics (TOG)34(6), 248:1– 248:16 (2015) 4, 5

Loper, M., Mahmood, N., Romero, J., Pons-Moll, G., Black, M.J.: SMPL: A skinned multi-person linear model. Transactions on Graphics (TOG)34(6), 248:1– 248:16 (2015) 4, 5

2015
[50]

In: International Conference on Learning Representations (ICLR) (2019) 10

Loshchilov, I., Hutter, F.: Decoupled weight decay regularization. In: International Conference on Learning Representations (ICLR) (2019) 10

2019
[51]

In: International Conference on Learning Representations (ICLR) (2022) 10, 11

Meng, C., He, Y., Song, Y., Song, J., Wu, J., Zhu, J., Ermon, S.: SDEdit: Guided image synthesis and editing with stochastic differential equations. In: International Conference on Learning Representations (ICLR) (2022) 10, 11

2022
[52]

In: Computer Vision and Pattern Recognition (CVPR)

Nam, H., Jung, D.S., Moon, G., Lee, K.M.: Joint reconstruction of 3D human and object via contact-based refinement transformer. In: Computer Vision and Pattern Recognition (CVPR). pp. 10218–10227 (2024) 4, 5, 11, 13

2024
[53]

Transactions on Ma- chine Learning Research (TMLR) (2024) 10

Oquab, M., Darcet, T., Moutakanni, T., Vo, H., Szafraniec, M., Khalidov, V., Fer- nandez, P., Haziza, D., Massa, F., El-Nouby, A., Assran, M., Ballas, N., Galuba, W., Howes, R., Huang, P.Y., Li, S.W., Misra, I., Rabbat, M., Sharma, V., Syn- naeve, G., Xu, H., Jegou, H., Mairal, J., Labatut, P., Joulin, A., Bojanowski, P.: DINOv2: Learning robust visual fe...

2024
[54]

In: Com- puter Vision and Pattern Recognition (CVPR)

Park, J.J., Florence, P.R., Straub, J., Newcombe, R.A., Lovegrove, S.: DeepSDF: Learning continuous signed distance functions for shape representation. In: Com- puter Vision and Pattern Recognition (CVPR). pp. 165–174 (2019) 4

2019
[55]

In: Computer Vision and Pattern Recognition (CVPR)

Paschalidou, D., Ulusoy, A.O., Geiger, A.: Superquadrics revisited: Learning 3D shape parsing beyond cuboids. In: Computer Vision and Pattern Recognition (CVPR). pp. 10344–10353 (2019) 4

2019
[56]

In: Interna- tional Conference on 3D Vision (3DV)

Patel, P., Black, M.: CameraHMR: Aligning people with perspective. In: Interna- tional Conference on 3D Vision (3DV). pp. 1562–1571 (2025) 4, 11, 12

2025
[57]

In: Computer Vision and Pattern Recognition (CVPR)

Pavlakos, G., Choutas, V., Ghorbani, N., Bolkart, T., Osman, A.A.A., Tzionas, D., Black, M.J.: Expressive body capture: 3D hands, face, and body from a single image. In: Computer Vision and Pattern Recognition (CVPR). pp. 10975–10985 (2019) 11

2019
[58]

In: International Conference on Computer Vision (ICCV)

Peebles, W., Xie, S.: Scalable diffusion models with transformers. In: International Conference on Computer Vision (ICCV). pp. 4172–4182 (2023) 8

2023
[59]

In: International Conference on Computer Vision (ICCV)

Petrov, I.A., Marin, R., Chibane, J., Pons-Moll, G.: TriDi: Trilateral diffusion of 3D humans, objects, and interactions. In: International Conference on Computer Vision (ICCV). pp. 5523–5535 (2025) 4, 8

2025
[60]

In: Computer Vision and Pattern Recognition (CVPR)

Qi, H., Zhao, C., Salzmann, M., Mathis, A.: HOISDF: Constraining 3D hand- object pose estimation with global signed distance fields. In: Computer Vision and Pattern Recognition (CVPR). pp. 10392–10402 (2024) 4

2024
[61]

In: Conference on Neural Information Processing Systems (NeurIPS) (2022) 7, 10

Qian, G., Li, Y., Peng, H., Mai, J., Hammoud, H., Elhoseiny, M., Ghanem, B.: PointNeXt: Revisiting pointnet++ with improved training and scaling strategies. In: Conference on Neural Information Processing Systems (NeurIPS) (2022) 7, 10

2022
[62]

In: International Conference on Learning Representations (ICLR) (2024) 4

Qian, G., Mai, J., Hamdi, A., Ren, J., Siarohin, A., Li, B., Lee, H.Y., Skorokhodov, I., Wonka, P., Tulyakov, S., Ghanem, B.: Magic123: One image to high-quality 3D object generation using both 2D and 3D diffusion priors. In: International Conference on Learning Representations (ICLR) (2024) 4

2024
[63]

Accelerating 3D Deep Learning with PyTorch3D

Ravi,N.,Reizenstein,J.,Novotny,D.,Gordon,T.,Lo,W.Y.,Johnson,J.,Gkioxari, G.: Accelerating 3D deep learning with PyTorch3D. arXiv:2007.08501 (2020) 10 20 D. Antić et al

work page internal anchor Pith review arXiv 2007
[64]

Transactions on Graphics (TOG)36(6), 245:1–245:17 (2017) 5, 7, 11

Romero, J., Tzionas, D., Black, M.J.: Embodied hands: Modeling and capturing hands and bodies together. Transactions on Graphics (TOG)36(6), 245:1–245:17 (2017) 5, 7, 11

2017
[65]

In: Conference on Neural Information Processing Systems (NeurIPS) (2024) 4

Sárándi, I., Pons-Moll, G.: Neural localizer fields for continuous 3D human pose and shape estimation. In: Conference on Neural Information Processing Systems (NeurIPS) (2024) 4

2024
[66]

Transactions on Graphics (TOG) 35(4), 139 (2016) 4

Savva, M., Chang, A.X., Hanrahan, P., Fisher, M., Nießner, M.: PiGraphs: Learn- ing interaction snapshots from observations. Transactions on Graphics (TOG) 35(4), 139 (2016) 4

2016
[67]

In: European Conference on Computer Vision (ECCV)

Shimada, S., Golyanik, V., Li, Z., Pérez, P., Xu, W., Theobalt, C.: HULC: 3D human motion capture with pose manifold sampling and dense contact guidance. In: European Conference on Computer Vision (ECCV). vol. 13682, pp. 516–533 (2022) 2

2022
[68]

In: Computer Vision and Pattern Recogni- tion (CVPR)

Taheri, O., Choutas, V., Black, M.J., Tzionas, D.: GOAL: Generating 4D whole- body motion for hand-object grasping. In: Computer Vision and Pattern Recogni- tion (CVPR). pp. 13263–13273 (2022) 2, 4

2022
[69]

In: European Conference on Computer Vision (ECCV)

Taheri, O., Ghorbani, N., Black, M.J., Tzionas, D.: GRAB: A dataset of whole- body human grasping of objects. In: European Conference on Computer Vision (ECCV). vol. 12349, pp. 581–600 (2020) 2, 4

2020
[70]

In: International Conference on 3D Vision (3DV)

Taheri, O., Zhou, Y., Tzionas, D., Zhou, Y., Ceylan, D., Pirk, S., Black, M.J.: GRIP: Generating interaction poses using spatial cues and latent consistency. In: International Conference on 3D Vision (3DV). pp. 933–943 (2024) 4

2024
[71]

SAM 3D: 3Dfy Anything in Images

Team, S.D., Chen, X., Chu, F.J., Gleize, P., Liang, K.J., Sax, A., Tang, H., Wang, W., Guo, M., Hardin, T., Li, X., Lin, A., Liu, J., Ma, Z., Sagar, A., Song, B., Wang, X., Yang, J., Zhang, B., Dollár, P., Gkioxari, G., Feiszli, M., Malik, J.: SAM 3D: 3Dfy anything in images. arXiv:2511.16624 (2025) 4, 5, 8, 10, 11, 12

work page internal anchor Pith review arXiv 2025
[72]

In: International Confer- ence on Computer Vision (ICCV)

Tripathi, S., Chatterjee, A., Passy, J., Yi, H., Tzionas, D., Black, M.J.: DECO: Dense estimation of 3D human-scene contact in the wild. In: International Confer- ence on Computer Vision (ICCV). pp. 7967–7979 (2023) 4, 5

2023
[73]

In: Computer Vision and Pattern Recognition (CVPR)

Tripathi, S., Müller, L., Huang, C.H.P., Omid, T., Black, M.J., Tzionas, D.: 3D human pose estimation via intuitive physics. In: Computer Vision and Pattern Recognition (CVPR). pp. 4713–4725 (2023) 4

2023
[74]

In: Computer Vision and Pattern Recognition (CVPR)

Wang, H., Sridhar, S., Huang, J., Valentin, J.P.C., Song, S., Guibas, L.: Normalized object coordinate space for category-level 6D object pose and size estimation. In: Computer Vision and Pattern Recognition (CVPR). pp. 2642–2651 (2019) 4

2019
[75]

In: European Conference on Computer Vision (ECCV)

Wang, N., Zhang, Y., Li, Z., Fu, Y., Liu, W., Jiang, Y.G.: Pixel2Mesh: Generating 3D mesh models from single RGB images. In: European Conference on Computer Vision (ECCV). vol. 11215, pp. 55–71 (2018) 4

2018
[76]

In: Computer Vision and Pattern Recognition (CVPR)

Wang,R.,Xu,S.,Dai,C.,Xiang,J.,Deng,Y.,Tong,X.,Yang,J.:MoGe:Unlocking accurate monocular geometry estimation for open-domain images with optimal training supervision. In: Computer Vision and Pattern Recognition (CVPR). pp. 5261–5271 (June 2025) 11, 12

2025
[77]

In: Computer Vision and Pattern Recognition (CVPR)

Wang, Z., Zheng, Q., Ma, S., Ye, M., Zhan, Y., Li, D.: End-to-end HOI reconstruc- tion transformer with graph-based encoding. In: Computer Vision and Pattern Recognition (CVPR). pp. 27706–27715 (2025) 5, 11, 13

2025
[78]

In: Computer Vision and Pattern Recognition (CVPR)

Wen, B., Huang, D., Zhang, Z., Zhou, J., Deng, J., Gong, J., Chen, Y., Ma, L., Li, Y.: Reconstructing in-the-wild open-vocabulary human-object interactions. In: Computer Vision and Pattern Recognition (CVPR). pp. 17426–17436 (2025) 3, 4, 5, 11, 12, 13, 14 LEXIS 21

2025
[79]

In: Computer Vision and Pattern Recognition (CVPR)

Wu, Z., Song, S., Khosla, A., Yu, F., Zhang, L., Tang, X., Xiao, J.: 3D shapenets: A deep representation for volumetric shapes. In: Computer Vision and Pattern Recognition (CVPR). pp. 1912–1920 (2015) 10

1912
[80]

In: Proceedings of Robotics: Science and Systems

Xiang, Y., Schmidt, T., Narayanan, V., Fox, D.: PoseCNN: A convolutional neu- ral network for 6D object pose estimation in cluttered scenes. In: Proceedings of Robotics: Science and Systems. Pittsburgh, Pennsylvania (June 2018) 4

2018

Showing first 80 references.