Pith. sign in

REVIEW 4 major objections 5 minor 2 cited by

MVHOI claims to be the first HOI video reenactment framework that transfers complex 3D manipulation dynamics — flips, spins, out-of-plane reorientations — onto novel objects by anchoring motion and appearance in a single 3D latent space.

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 ↗

Using a 3D foundation model to produce viewpoint-aware anchors plus multi-view reference textures enables realistic human-object-interaction reenactment with large out-of-plane rotations.

T0 review reviewed 2026-08-04 challenge →

load-bearing objection A well-engineered two-stage HOI reenactment system with real gains, but the evaluation is too thin and the object-agnostic motion assumption untested to fully back the headline claim. the 4 major comments →

arxiv 2603.14686 v2 pith:R4SSMXLS submitted 2026-03-16 cs.CV cs.AI

MVHOI: Bridge Multi-view Condition to Complex Human-Object Interaction Video Reenactment via 3D Foundation Model

Jinguang Tong , Jinbo Wu , Kaisiyuan Wang , Zhelun Shen , Xuan Huang , Mochu Xiang , Xuesong Li , Yingying Li
show 7 more authors
Haocheng Feng Chen Zhao Hang Zhou Wei He Chuong Nguyen Jingdong Wang Hongdong Li
This is my paper
classification cs.CV cs.AI
keywords human-object interactionvideo reenactment3D foundation modelmulti-view referencediffusion transformerobject swapview synthesiscontrollable video generation
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

This paper argues that the fundamental limitation of prior HOI reenactment methods is their reliance on sparse 2D controls (bounding boxes, keypoints) and monocular references, which cannot represent the view-dependent geometry of an object during rotation. The proposed two-stage framework, MVHOI, claims to solve this by embedding a 3D foundation model as a Unified Object Anchor: the anchor consumes frozen motion embeddings from the source video and multi-view references of the target object, and autoregressively predicts coarse, view-consistent object frames. These frames then guide a diffusion-transformer video generator, which also uses an attention bias derived from the anchor's internal cross-view attention to retrieve the correct reference texture per viewpoint. If the claims hold, HOI reenactment would no longer be restricted to image-plane translations and could handle rapid rotations, severe hand-object occlusion, and long videos without cumulative drift.

Core claim

The central claim is that a feed-forward 3D foundation model, used as a Unified Object Anchor (UOA), can consolidate sparse multi-view references and implicit motion latents into a view-invariant latent repository, and that querying this anchor over time yields a sequence of coarse object states that are geometrically consistent with the source interaction. This replaces explicit pose estimation and 2D proxies. The paper further claims that the UOA's internal attention map, extracted at inference time, encodes a view-selection signal that can be added as a logit-level bias to a DiT generator's self-attention, thereby steering it to attend to the correct reference view. Together, these two me

What carries the argument

The Unified Object Anchor (UOA): a transformer backbone (initialized from a large 3D foundation model) that takes the current target-object frame plus multi-view reference images, modulates the frame tokens with scale-and-shift from a frozen motion embedding, and predicts the next object frame through a dedicated RGB head. Its role is to provide a deterministic view-querying pathway in latent space, turning stochastic synthesis into a structured anchor-following task. The same anchor's cross-view attention map is reused as a soft bias in the video generator's context adapter, directly linking geometric reasoning to appearance retrieval.

Load-bearing premise

The whole transfer pipeline rests on the premise that the frozen motion extractor's latent embeddings describe the physical dynamics of the interaction and are independent of the source object's identity, category, or appearance.

What would settle it

Take a single source interaction (e.g., a hand flipping a phone 360 degrees) and run it with source objects of very different visual and physical properties (a rigid mug, a soft plush toy, a transparent glass). If the extracted motion embeddings differ measurably across source objects for identical hand kinematics, or if cross-reenactment quality on a novel rigid target drops sharply for one of these source videos in a user study, the object-agnostic motion assumption is undermined. A second probe: replace the UOA's coarse anchors with a simple per-frame bounding-box trajectory under the same

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

If this is right

  • Out-of-plane object rotations and flips become tractable control signals, so reenactment can follow a hand twirling a bottle or flipping a phone rather than only sliding it on a table.
  • Object swapping across categories (e.g., replacing a mug with a stapler) can preserve motion plausibility because the anchor supplies view-consistent geometry while the generator fetches texture from the correct reference view.
  • Long videos avoid drift because the cross-iterative loop re-initializes each segment from the previous refined output, not from blurred intermediate anchors.
  • The recipe — a 3D foundation model as a latent anchor plus an attention-bias steering mechanism — offers a general pattern for injecting 3D awareness into existing video diffusion models without retraining them fully.

Where Pith is reading between the lines

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

  • The authors freeze the motion extractor and never test whether its descriptors are truly object-agnostic; if the same motion on a rigid mug and a plush toy produces different embeddings, the cross-category transfer claim weakens. A simple invariance test would settle this.
  • Since the UOA is trained on rendered 3D objects, its anchor quality may degrade on transparent, reflective, or deformable objects; if so, the framework's claim of 'complex non-planar dynamics' may hold only for rigid, diffuse objects in the training distribution.
  • The attention-bias injection could be applied to other editing tasks — e.g., viewpoint control in object-centric video generation, or maintaining identity across regenerations — as a drop-in prior.
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

4 major / 5 minor

Summary. The paper presents MVHOI, a two-stage framework for human-object interaction (HOI) video reenactment. Stage I extracts motion embeddings from a source video using a frozen DisMo motion extractor, then conditions a DepthAnything3-based Unified Object Anchor (UOA) on those embeddings and multi-view references of a target object to autoregressively predict coarse object-anchor frames. Stage II uses a DiT-based video generator (initialized from Wan2.1-T2V-14B, with a VACE-style context adapter) that ingests the coarse anchors and multi-view references, and applies an inference-time attention bias derived from intermediate UOA attention maps to reduce reference-view confusion. A cross-iterative inference scheme alternates between coarse anchor prediction and video refinement for long sequences. The paper claims to be the first HOI reenactment framework to handle complex non-planar dynamics, and reports quantitative improvements over MimicMotion, VACE, HuMo, and HunyuanCustom on self- and cross-reenactment, plus an ablation of the coarse-guidance and attention-enhancement components.

Significance. If the reported results are reliable, the proposed direction is significant: leveraging 3D foundation models as an intermediate anchor for transferring object dynamics is a plausible and timely alternative to explicit 6D pose estimation or 2D keypoint control, and the two-stage design is well-motivated. The paper is largely free of circularity: the performance claims are empirical, the pipeline is described in enough detail to be implemented, and the ablation in Table 4 suggests that both the coarse guidance and the attention bias contribute to the final result. The reported metric gaps over VACE (e.g., self-reenactment FID 17.9 vs 47.4) are large, which gives the central claim initial plausibility. However, the evidence is weakened by the small, self-selected public evaluation set, the proprietary training and test data, the absence of the closest baseline ReHold, and the untested assumption that DisMo motion embeddings are object-agnostic. No code or data is released, which further limits verifiability.

major comments (4)
  1. [Sec. 3.2, Eq. (1)] The cross-object transfer mechanism relies on the frozen DisMo motion extractor producing descriptors that are object-agnostic. The sentence “This latent representation serves as a compact descriptor of the underlying physical dynamics” is asserted but never probed. If m_t encodes source-object shape, texture, or category, the UOA’s predicted anchors are contaminated before Stage II begins, regardless of the strength of the video generator. I request a dedicated experiment: e.g., train a linear probe to predict source-object identity from m_t; stratify cross-reenactment results by source–target category similarity; or compare UOA outputs when the same target object is driven by motions from different source objects. Without this, the central “object-to-object” claim rests on an untested hidden condition.
  2. [Sec. 4.2, Table 1] The main quantitative claim is supported by a weak evaluation protocol. Cross-reenactment is evaluated on 29 self-selected samples from the AnchorCrafter training split, and the self-collected 100-video set has no stated split or selection criterion. No error bars, confidence intervals, or significance tests are reported for any metric in Tables 1–4. The headline FID/FVD gaps (e.g., 17.9 vs 47.4 in self-reenactment) could be affected by sample selection or noise. Add bootstrap confidence intervals or per-sample distributions, use a held-out test set, and specify the selection protocol for both the 29 AnchorCrafter samples and the 100 self-collected videos.
  3. [Sec. 2.2 / Sec. 4.3] ReHold [7] is identified in Related Work as the closest HOI reenactment baseline, but it is never included in the experiments. Without this comparison, the claim to be “the first HOI video reenactment framework capable of synthesizing complex non-planar dynamics” and to “transcend the limitations of 2D translational models” is not empirically substantiated relative to the most relevant prior work. Please add ReHold (and, if feasible, DreamActor-H1 [30]) to Tables 1 and 3, or explicitly justify their exclusion.
  4. [Sec. 3.5, Table 3] Cross-iterative long-video inference is a stated contribution, but the experiments only compare the full model against external baselines. There is no ablation comparing the full model with and without the cross-iterative loop, so the specific claim that the strategy “effectively mitigates drift” is not directly tested. Add an ablation where long sequences are generated without re-initializing from refined video segments (e.g., naive autoregressive rollout from previous coarse anchors) to isolate the benefit of the proposed mechanism.
minor comments (5)
  1. [Sec. 3.3, Eq. (3)] The attention-bias strength α is not reported, nor is the designated UOA layer from which the attention map w is extracted. These details are needed for reproducibility. Similarly, the HOI loss weight β in Eq. (6) is not specified.
  2. [Sec. 3.2] The initialization O_0 for the autoregressive rollout in cross-reenactment is not specified. It should be clarified how the first target-object frame is obtained from the multi-view references and aligned with the source video’s initial hand state.
  3. [Sec. 4.3, Table 2] The Stage I comparison against DisMo conflates the UOA architecture with the addition of multi-view reference conditioning, since DisMo does not receive multi-view references. An ablation with the same reference inputs but without the UOA mechanism would better isolate the contribution of the proposed module.
  4. [Sec. 4.2] The metrics O-CLIP, MC, VQ, and HR are not defined in the main text; the paper refers to an Appendix that is not included in the submitted version. Please define them or include the appendix.
  5. [Throughout] There are numerous typos and formatting artifacts, including “shwon”, “structed”, “Ojbect”, “adpots”, “desin”, “conherence”, “freezen”, and “achieves”. References [17] and [18] are duplicates of the same Depth Anything 3 paper. The final “Received 20 February 2007; revised 12 March 2009; accepted 5 June 2009” line appears to be a template artifact and should be removed.

Circularity Check

0 steps flagged

No significant circularity: the framework's outputs are not defined as its inputs, and the only group self-citation (ReHold) is in related work and is not load-bearing.

full rationale

MVHOI's derivation chain is empirical rather than tautological. Stage I trains a UOA on rendered Objaverse data with photometric losses, and Stage II trains a DiT generator on proxy guidance augmented to mimic UOA degradation; the final inference passes UOA coarse anchors and an attention-map bias into the video generator. Neither equation reduces the predicted output to a fitted input: the coarse anchors are a learned prediction, and the attention bias is an intermediate activation, not a quantity optimized to match the evaluation metrics. The central contribution is benchmarked against external baselines (MimicMotion, VACE, HunyuanCustom, HuMo, DisMo), so the performance claims are self-contained and not forced by construction. The only related self-citation is ReHold [7], which shares authors with this paper; it appears in the related-work discussion, not as evidence for MVHOI's claims or as a forbidden 'uniqueness' theorem. The paper's untested assumption that DisMo motion embeddings are object-agnostic is a correctness risk, but an unverified premise is not circularity unless the conclusion is already embedded in the premise. The mild internal dependence that the Stage-II attention bias comes from the same UOA that produces Stage-I anchors is a design coupling, not a tautology: the attention map is a distinct output and is not fitted to make the final video metrics succeed by definition. Overall, the derivation is self-contained against external benchmarks, and no specific reduction from Eq. X to Eq. Y or fitted-parameter-renamed-as-prediction can be exhibited. Score 2 reflects the minor non-load-bearing self-citation, not actual circularity.

Axiom & Free-Parameter Ledger

4 free parameters · 6 axioms · 0 invented entities

The central claim rests on several domain assumptions about pretrained components that are not independently validated: DisMo motion descriptors are object-agnostic; DepthAnything3 features can act as a navigable object anchor; UOA attention indicates true viewpoint; Stage II's augmentation mimics real UOA outputs; synthetic Objaverse training transfers to real video. Hand-set hyperparameters (Δt, λ, α, β) affect results, though none are fit to test data in a way that would make the performance claims circular.

free parameters (4)
  • temporal stride Δt = 4
    Set to 4 in all experiments (Sec. 3.2, Eq. 1); hand-chosen, affects motion descriptor granularity and UOA rollout spacing.
  • loss weights λ1, λ2, λ3 = 1.0, 0.1, 0.1
    Stage I training weights (Sec. 3.4, Eq. 4); chosen empirically. Central claim depends on tuned UOA outputs.
  • attention bias strength α = not reported
    In Eq. 3, α controls the effect of the geometry-aware attention bias; no value or sensitivity analysis provided. This is a hand-tuned inference-time parameter.
  • HOI loss weight β = not reported
    Introduced in Eq. 6 to balance HOI-region loss; no value given. Affects the relative emphasis on interaction regions.
axioms (6)
  • domain assumption DisMo motion embeddings are object-agnostic and transferable across object identities
    Sec 3.2 freezes the DisMo motion extractor and relies on it to encode dynamics independent of the source object; the paper gives no probe of this property beyond qualitative results.
  • domain assumption DepthAnything3's feature aggregator functions as a unified object anchor that can be navigated with motion-conditioned scale/shift
    Sec 3.2 instantiates UOA on DA3-Large; assumes its latent features align views enough for view-querying, without explicit 3D reconstruction validation.
  • domain assumption UOA attention maps correlate with true viewpoint of the object
    Sec 3.3 uses intermediate attention w as a geometric prior for view selection; evidenced only by one visualization (Fig. 4) and an ablation, not by measured viewpoint accuracy.
  • domain assumption Proxy-guidance augmentation in Stage II training mimics the actual UOA output distribution
    Sec 3.3 trains the video generator on original object crops with augmentation instead of real UOA outputs; if the augmentation gap is large, the generator fails on true coarse anchors.
  • domain assumption Objaverse synthetic renders transfer to real driving videos
    Stage I is trained on 40k Objaverse objects with synthetic motion frames, then evaluated on real videos; no explicit domain adaptation is reported.
  • standard math Flow-matching and DiT attention softmax are valid background machinery
    Sec 3.3-3.4; inherited from Wan2.1 and VACE.

reviewed 2026-08-04 · how reviews work

0 comments
Cite this review

Pith. "Pith review of MVHOI: Bridge Multi-view Condition to Complex Human-Object Interaction Video Reenactment via 3D Foundation Model." pith.science (2026). https://pith.science/paper/R4SSMXLS

@misc{pith2026260314686,
  author       = {Pith},
  title        = {Pith review of: MVHOI: Bridge Multi-view Condition to Complex Human-Object Interaction Video Reenactment via 3D Foundation Model},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/R4SSMXLS}},
  note         = {Machine review of arXiv:2603.14686}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Human-Object Interaction (HOI) video reenactment aims to transfer the interaction dynamics of a source video to a novel target object while preserving realistic hand-object coordination. Existing methods typically rely on sparse 2D motion controls and monocular references, which are insufficient for complex out-of-plane motion and large viewpoint changes. We present MVHOI, a two-stage framework combining implicit motion extraction, 3D-aware multi-view reasoning, and video generation. In the first stage, a motion extractor encodes object dynamics into implicit motion descriptors. Conditioned on these descriptors, our Motion-Driven Object Prior (MDOP) module queries a 3D foundation model over multi-view references of the target object and autoregressively predicts coarse object anchors, a sequence of images that track the object's evolving orientation and appearance under the source motion without any explicit pose estimation. In the second stage, a DiT-based video generation model uses these anchors as structural guidance and the multi-view references as appearance guidance. We further reuse cross-view attention from MDOP as a soft attention bias to reduce reference-view confusion. For long videos, a cross-iterative inference strategy refreshes subsequent object priors using refined video outputs. Experiments demonstrate consistent improvements over state-of-the-art methods in object fidelity, motion consistency, visual quality, and interaction realism.

Figures

Figures reproduced from arXiv: 2603.14686 by Chen Zhao, Chuong Nguyen, Hang Zhou, Haocheng Feng, Hongdong Li, Jinbo Wu, Jingdong Wang, Jinguang Tong, Kaisiyuan Wang, Mochu Xiang, Wei He, Xuan Huang, Xuesong Li, Yingying Li, Zhelun Shen.

Figure 1
Figure 1. Figure 1: MVHOI enables high-fidelity Human-Object Interaction (HOI) video reenactment by leveraging multi-view reference [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Overview of MVHOI. Stage I extracts motion features from the source interaction and combines them with multi-view [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Qualitative comparison with state-of-the-art methods. Compared to MimicMotion [ [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Visualization of the attention map. This heat map [PITH_FULL_IMAGE:figures/full_fig_p006_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Qualitative results for 3D-aware motion transfer. [PITH_FULL_IMAGE:figures/full_fig_p007_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Qualitative ablation results. CRG improves motion [PITH_FULL_IMAGE:figures/full_fig_p008_6.png] view at source ↗

discussion (0)

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

Forward citations

Cited by 2 Pith papers

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

  1. VistaRef: Boosting Visual Spatial Orientation Awareness for Pointing-to-Object Detection

    cs.CV 2026-06 unverdicted novelty 5.0

    VistaRef improves pointing-to-object detection accuracy by 14 points via local hand entity modeling, geometric ray modeling, and an orientation-consistent alignment loss.

  2. Controllable Video Object Insertion via Multiview Priors

    cs.CV 2026-04 unverdicted novelty 5.0

    A multi-view prior-based framework for video object insertion that uses dual-path conditioning and an integration-aware consistency module to improve appearance stability and occlusion handling.

Reference graph

Works this paper leans on

47 extracted references · 21 linked inside Pith · cited by 2 Pith papers

  1. [1]

    Andreas Blattmann, Tim Dockhorn, Sumith Kulal, Daniel Mendelevitch, Maciej Kilian, Dominik Lorenz, Yam Levi, Zion English, Vikram Voleti, Adam Letts, et al. 2023. Stable video diffusion: Scaling latent video diffusion models to large datasets.arXiv preprint arXiv:2311.15127(2023)

    Pith/arXiv arXiv 2023

  2. [2]

    Tim Brooks, Bill Peebles, Connor Holmes, Will DePue, Yufei Guo, Li Jing, David Schnurr, Joe Taylor, Troy Luhman, Eric Luhman, Clarence Ng, Ricky Wang, and Aditya Ramesh. 2024. Video generation models as world simulators. (2024). https://openai.com/research/video-generation-models-as-world-simulators

    2024

  3. [3]

    Liyang Chen, Tianxiang Ma, Jiawei Liu, Bingchuan Li, Zhuowei Chen, Lijie Liu, Xu He, Gen Li, Qian He, and Zhiyong Wu. 2025. Humo: Human-centric video generation via collaborative multi-modal conditioning.arXiv preprint arXiv:2509.08519(2025)

    Pith/arXiv arXiv 2025

  4. [4]

    Gang Cheng, Xin Gao, Li Hu, Siqi Hu, Mingyang Huang, Chaonan Ji, Ju Li, Dechao Meng, Jinwei Qi, Penchong Qiao, et al . 2025. Wan-animate: Unified character animation and replacement with holistic replication.arXiv preprint arXiv:2509.14055(2025)

    arXiv 2025

  5. [5]

    Ruihang Chu, Yefei He, Zhekai Chen, Shiwei Zhang, Xiaogang Xu, Bin Xia, Dingdong Wang, Hongwei Yi, Xihui Liu, Hengshuang Zhao, et al. 2025. Wan- move: Motion-controllable video generation via latent trajectory guidance.arXiv preprint arXiv:2512.08765(2025)

    arXiv 2025

  6. [6]

    Matt Deitke, Dustin Schwenk, Jordi Salvador, Luca Weihs, Oscar Michel, Eli VanderBilt, Ludwig Schmidt, Kiana Ehsani, Aniruddha Kembhavi, and Ali Farhadi

  7. [7]

    Yingying Fan, Quanwei Yang, Kaisiyuan Wang, Hang Zhou, Yingying Li, Haocheng Feng, Errui Ding, Yu Wu, and Jingdong Wang. 2025. Re-hold: Video hand object interaction reenactment via adaptive layout-instructed diffusion model. InProceedings of the Computer Vision and Pattern Recognition Conference. 17550–17560

    2025

  8. [8]

    Jiazhi Guan, Zhiliang Xu, Hang Zhou, Kaisiyuan Wang, Shengyi He, Zhanwang Zhang, Borong Liang, Haocheng Feng, Errui Ding, Jingtuo Liu, Jingdong Wang, Youjian Zhao, and Ziwei Liu. 2024. Resyncer: Rewiring Style-based Generator for Unified Audio-Visually Synced Facial Performer. InECCV

    2024

  9. [9]

    Yuwei Guo, Ceyuan Yang, Anyi Rao, Zhengyang Liang, Yaohui Wang, Yu Qiao, Maneesh Agrawala, Dahua Lin, and Bo Dai. 2023. Animatediff: Animate your personalized text-to-image diffusion models without specific tuning.arXiv preprint arXiv:2307.04725(2023)

    Pith/arXiv arXiv 2023

  10. [10]

    Amir Hertz, Ron Mokady, Jay Tenenbaum, Kfir Aberman, Yael Pritch, and Daniel Cohen-Or. 2022. Prompt-to-prompt image editing with cross attention con- trol.(2022).URL https://arxiv. org/abs/2208.016263 (2022)

    Pith/arXiv arXiv 2022

  11. [11]

    Li Hu, Xin Gao, Peng Zhang, Ke Sun, Bang Zhang, and Liefeng Bo. 2023. Ani- mate anyone: Consistent and controllable image-to-video synthesis for character animation.arXiv preprint arXiv:2311.17117(2023)

    Pith/arXiv arXiv 2023

  12. [12]

    Teng Hu, Zhentao Yu, Zhengguang Zhou, Sen Liang, Yuan Zhou, Qin Lin, and Qinglin Lu. 2025. Hunyuancustom: A multimodal-driven architecture for cus- tomized video generation.arXiv preprint arXiv:2505.04512(2025)

    Pith/arXiv arXiv 2025

  13. [13]

    Ziyao Huang, Zixiang Zhou, Juan Cao, Yifeng Ma, Yi Chen, Zejing Rao, Zhiyong Xu, Hongmei Wang, Qin Lin, Yuan Zhou, et al. 2025. HunyuanVideo-HOMA: Generic Human-Object Interaction in Multimodal Driven Human Animation. arXiv preprint arXiv:2506.08797(2025)

    Pith/arXiv arXiv 2025

  14. [14]

    Tao Jiang, Xinchen Xie, and Yining Li. 2024. RTMW: Real-time multi-person 2D and 3D whole-body pose estimation.arXiv preprint arXiv:2407.08634(2024)

    Pith/arXiv arXiv 2024

  15. [15]

    Zeyinzi Jiang, Zhen Han, Chaojie Mao, Jingfeng Zhang, Yulin Pan, and Yu Liu

  16. [16]

    Weijie Kong, Qi Tian, Zijian Zhang, Rox Min, Zuozhuo Dai, Jin Zhou, Jiangfeng Xiong, Xin Li, Bo Wu, Jianwei Zhang, et al. 2024. Hunyuanvideo: A systematic framework for large video generative models.arXiv preprint arXiv:2412.03603 (2024)

    Pith/arXiv arXiv 2024

  17. [18]

    Haotong Lin, Sili Chen, Junhao Liew, Donny Y Chen, Zhenyu Li, Guang Shi, Jiashi Feng, and Bingyi Kang. 2025. Depth anything 3: Recovering the visual space from any views.arXiv preprint arXiv:2511.10647(2025)

    Pith/arXiv arXiv 2025

  18. [19]

    Yaron Lipman, Ricky TQ Chen, Heli Ben-Hamu, Maximilian Nickel, and Matt Le

  19. [20]

    Lijie Liu, Tianxiang Ma, Bingchuan Li, Zhuowei Chen, Jiawei Liu, Gen Li, Siyu Zhou, Qian He, and Xinglong Wu. 2025. Phantom: Subject-consistent video generation via cross-modal alignment.arXiv preprint arXiv:2502.11079(2025)

    Pith/arXiv arXiv 2025

  20. [21]

    Xingchao Liu, Chengyue Gong, and Qiang Liu. 2022. Flow straight and fast: Learning to generate and transfer data with rectified flow.arXiv preprint arXiv:2209.03003(2022)

    Pith/arXiv arXiv 2022

  21. [22]

    Yuxuan Luo, Zhengkun Rong, Lizhen Wang, Longhao Zhang, and Tianshu Hu

  22. [23]

    William Peebles and Saining Xie. 2023. Scalable diffusion models with transform- ers. InProceedings of the IEEE/CVF international conference on computer vision. 4195–4205

    2023

  23. [24]

    Ethan Perez, Florian Strub, Harm De Vries, Vincent Dumoulin, and Aaron Courville. 2018. Film: Visual reasoning with a general conditioning layer. In Proceedings of the AAAI conference on artificial intelligence, Vol. 32

    2018

  24. [25]

    Nikhila Ravi, Valentin Gabeur, Yuan-Ting Hu, Ronghang Hu, Chaitanya Ryali, Tengyu Ma, Haitham Khedr, Roman Rädle, Chloe Rolland, Laura Gustafson, et al. 2024. Sam 2: Segment anything in images and videos.arXiv preprint arXiv:2408.00714(2024)

    Pith/arXiv arXiv 2024

  25. [26]

    InProceedings of the IEEE/CVF International Conference on Computer Vision

    Dreamactor-m1: Holistic, expressive and robust human image animation with hybrid guidance. InProceedings of the IEEE/CVF International Conference on Computer Vision. 11036–11046

  26. [27]

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Björn Ommer. 2022. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 10684–10695

    2022

  27. [28]

    Team Wan, Ang Wang, Baole Ai, Bin Wen, Chaojie Mao, Chen-Wei Xie, Di Chen, Feiwu Yu, Haiming Zhao, Jianxiao Yang, Jianyuan Zeng, Jiayu Wang, Jingfeng Zhang, Jingren Zhou, Jinkai Wang, Jixuan Chen, Kai Zhu, Kang Zhao, Keyu Yan, Lianghua Huang, Mengyang Feng, Ningyi Zhang, Pandeng Li, Pingyu Wu, Ruihang Chu, Ruili Feng, Shiwei Zhang, Siyang Sun, Tao Fang, T...

    Pith/arXiv arXiv 2025

  28. [29]

    Jianyuan Wang, Minghao Chen, Nikita Karaev, Andrea Vedaldi, Christian Rup- precht, and David Novotny. 2025. Vggt: Visual geometry grounded transformer. InProceedings of the Computer Vision and Pattern Recognition Conference. 5294– 5306

    2025

  29. [30]

    Thomas Ressler-Antal, Frank Fundel, Malek Ben Alaya, Stefan Andreas Baumann, Felix Krause, Ming Gui, and Björn Ommer. 2025. DisMo: Disentangled Motion Representations for Open-World Motion Transfer.arXiv preprint arXiv:2511.23428 (2025)

    arXiv 2025

  30. [31]

    Shuzhe Wang, Vincent Leroy, Yohann Cabon, Boris Chidlovskii, and Jerome Revaud. 2024. Dust3r: Geometric 3d vision made easy. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 20697–20709

    2024

  31. [32]

    Yifan Wang, Jianjun Zhou, Haoyi Zhu, Wenzheng Chang, Yang Zhou, Zizun Li, Junyi Chen, Jiangmiao Pang, Chunhua Shen, and Tong He. 2025. Permutation- Equivariant Visual Geometry Learning.arXiv preprint arXiv:2507.13347(2025)

    Pith/arXiv arXiv 2025

  32. [33]

    Jay Zhangjie Wu, Yixiao Ge, Xintao Wang, Stan Weixian Lei, Yuchao Gu, Yufei Shi, Wynne Hsu, Ying Shan, Xiaohu Qie, and Mike Zheng Shou. 2023. Tune-a- video: One-shot tuning of image diffusion models for text-to-video generation. InProceedings of the IEEE/CVF international conference on computer vision. 7623– 7633

    2023

  33. [34]

    Lizhen Wang, Zhurong Xia, Tianshu Hu, Pengrui Wang, Pengfei Wei, Zerong Zheng, Ming Zhou, Yuan Zhang, and Mingyuan Gao. 2025. Dreamactor-h1: High- fidelity human-product demonstration video generation via motion-designed diffusion transformers.arXiv preprint arXiv:2506.10568(2025)

    Pith/arXiv arXiv 2025

  34. [35]

    Zihui Xue, Mi Luo, Changan Chen, and Kristen Grauman. 2024. HOI-Swap: Swapping Objects in Videos with Hand-Object Interaction Awareness.NeurIPS (2024)

    2024

  35. [36]

    Zihui Sherry Xue, Romy Luo, Changan Chen, and Kristen Grauman. 2024. Hoi- swap: Swapping objects in videos with hand-object interaction awareness.Ad- vances in Neural Information Processing Systems37 (2024), 77132–77164

    2024

  36. [37]

    Lihe Yang, Bingyi Kang, Zilong Huang, Xiaogang Xu, Jiashi Feng, and Heng- shuang Zhao. 2024. Depth anything: Unleashing the power of large-scale un- labeled data. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition. 10371–10381

    2024

  37. [38]

    Ziyi Xu, Ziyao Huang, Juan Cao, Yong Zhang, Xiaodong Cun, Qing Shuai, Yuchen Wang, Linchao Bao, Jintao Li, and Fan Tang. 2024. AnchorCrafter: Animate Cyber- Anchors Saling Your Products via Human-Object Interacting Video Generation. arXiv preprint arXiv:2411.17383(2024). Conference acronym ’XX, June 03–05, 2018, Woodstock, NY Trovato et al

    Pith/arXiv arXiv 2024

  38. [39]

    Zhuoyi Yang, Jiayan Teng, Wendi Zheng, Ming Ding, Shiyu Huang, Jiazheng Xu, Yuanming Yang, Wenyi Hong, Xiaohan Zhang, Guanyu Feng, et al . 2024. CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer.arXiv preprint arXiv:2408.06072(2024)

    Pith/arXiv arXiv 2024

  39. [40]

    Yufei Ye, Poorvi Hebbar, Abhinav Gupta, and Shubham Tulsiani. 2023. Diffusion- guided reconstruction of everyday hand-object interaction clips. InCVPR

    2023

  40. [41]

    Tianyu Yu, Zefan Wang, Chongyi Wang, Fuwei Huang, Wenshuo Ma, Zhihui He, Tianchi Cai, Weize Chen, Yuxiang Huang, Yuanqian Zhao, et al. 2025. Minicpm-v 4.5: Cooking efficient mllms via architecture, data, and training recipe.arXiv preprint arXiv:2509.18154(2025)

    Pith/arXiv arXiv 2025

  41. [42]

    Lihe Yang, Bingyi Kang, Zilong Huang, Zhen Zhao, Xiaogang Xu, Jiashi Feng, and Hengshuang Zhao. 2024. Depth anything v2.Advances in Neural Information Processing Systems37 (2024), 21875–21911

    2024

  42. [43]

    Zhenhao Zhang, Ye Shi, Lingxiao Yang, Suting Ni, Qi Ye, and Jingya Wang. 2025. OpenHOI: Open-World Hand-Object Interaction Synthesis with Multimodal Large Language Model.arXiv preprint arXiv:2505.18947(2025)

    arXiv 2025

  43. [44]

    Zhenxing Zhang, Jiayan Teng, Zhuoyi Yang, Tiankun Cao, Cheng Wang, Xiaotao Gu, Jie Tang, Dan Guo, and Meng Wang. 2025. Kaleido: Open-Sourced Multi- Subject Reference Video Generation Model.arXiv preprint arXiv:2510.18573 (2025). Received 20 February 2007; revised 12 March 2009; accepted 5 June 2009

    arXiv 2025

  44. [46]

    Yuang Zhang, Jiaxi Gu, Li-Wen Wang, Han Wang, Junqi Cheng, Yuefeng Zhu, and Fangyuan Zou. 2025. MimicMotion: High-Quality Human Motion Video Generation with Confidence-aware Pose Guidance. InInternational Conference on Machine Learning

    2025

  45. [2022]

    Flow matching for generative modeling.arXiv preprint arXiv:2210.02747 (2022)

    Pith/arXiv arXiv 2022

  46. [2023]

    InProceedings of the IEEE/CVF conference on computer vision and pattern recognition

    Objaverse: A universe of annotated 3d objects. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition. 13142–13153

  47. [2025]

    Vace: All-in-one video creation and editing.arXiv preprint arXiv:2503.07598 (2025)

    Pith/arXiv arXiv 2025

This paper was first reviewed by deepseek-v4-flash on August 4, 2026.