Pith. sign in

REVIEW 5 major objections 5 minor 62 references

StateFlow: Building, Evolving, and Accessing 3D World States for Previsualization

T0 review · 5 major / 5 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read StateFlow claims that previsualization is best modeled as a persistent 3D world state, not as one-shot image or video generation.

desk verdict A genuinely useful state-centric reframing for previs, but the 'evolve' pillar is under-tested and the evaluation needs cleaner separation from the VLM it uses. read the letter →

arxiv 2608.12314 v1 pith:QIEXYAXK submitted 2026-08-12 cs.CV

Yuyang Yin , Zixiang Li , Longxuan Deng , Hongkai Li , Shifang Zhao , Junnan Liu , Weirong Huang , Mengyu Wang
show 6 more authors
Tianxiao Fu Yikai Wang Peng-Shuai Wang Xiaojie Jin Yao Zhao Yunchao Wei
This is my paper · ORCID
classification cs.CV MSC 68T45
keywords previsualization3Dworldstateobject-centricrepresentationdual-viewinitializationcameraplanningrender-feedbackreflectionintent-guidedtransitiongameprototyping
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper argues that previsualization, the iterative planning of scenes, cameras, and motion before final production, is poorly served by one-shot generators that entangle scene content, layout, motion, and camera in a single prompt and expose no persistent state to edit. Its proposal, StateFlow, maintains an explicit object-centric 3D world state, written as a set of object entities each with geometry, placement, and semantic attributes, and updates that state through three stages: construction from dual-view references with conflict resolution, evolution through structured transitions that preserve world memory, and access through camera trajectories refined by rendered feedback. The point of the state is reuse: different frames are local modifications or recombinations of shared geometry, pose, and semantics, so creators can iterate without regenerating the whole world. The paper reports that this state-centric approach outperforms video-generation and scene-generation baselines on consistency, controllability, and previsualization usefulness in both automatic metrics and user studies.

What carries the argument

The load-bearing object is the structured 3D world state, written as a set of object entities, each a triple of geometry, placement, and semantic attributes, together with three operations on it. State construction uses prior-guided conflict-aware dual-view initialization: a front view supplies appearance and asset crops, a bird's-eye view supplies ground-plane layout, and a vision-language model resolves count and placement conflicts before an inference-time objective refines the 3D boxes. State evolution applies intent-guided structured state transition: a vision-language model reads the state table and emits a compact transition plan, which is applied selectively over geometry, placement, or semantics. State access performs world-state camera planning with render-feedback reflection: the vision-language model proposes a trajectory, rendering exposes visibility, framing, and collision failures, and local repairs are scored and selected. The state itself is what carries reuse, because geometry, placement, and semantics are separate fields that can be updated independently.

What would settle it

Take a prompt in which the bird's-eye view implies an object count that the front view contradicts, for example text says 'three towers', the front view shows two, and the BEV shows three. If the vision-language model's conflict resolution keeps a hallucinated tower or drops one clearly visible in both views, and the lightweight box-optimization objective does not correct it, then the claim that StateFlow reconciles cross-view conflicts and builds physically plausible worlds is falsified. A second check is to render the constructed world from an unseen azimuth and look for floating, interpenetrating, or boundary-violating objects; any such artifact that persists after the physical-refinement term demonstrates that VLM semantic priors alone cannot guarantee physical plausibility.

Watch

Extended reading notes

Core claim

The central claim, stated on the paper's own terms, is that previsualization should be formulated as persistent 3D world-state modeling rather than one-shot visual synthesis. Concretely, a world is represented at time t as a set of object entities, each with geometry for rendering and physical reasoning, placement for spatial layout, and semantic attributes for editable state. Construction initializes this world from user conditions by combining a front view that supplies appearance and asset crops with a bird's-eye view that supplies ground-plane layout, using a vision-language model to resolve cross-view conflicts before a lightweight optimization refines 3D boxes. Evolution translates user intent into compact transitions over the state table, so edits are selective and memory-aware rather than full regenerations. Access plans camera trajectories that are proposed semantically and then repaired through rendered feedback. Against video-generation baselines and scene-generation baselines, StateFlow reports the best average VBench score and the best user-study and MLLM ratings for spatial and identity consistency, layout plausibility, and previsualization usefulness, with ablations showing the contribution of BEV grounding, conflict resolution, and render-feedback reflection.

Load-bearing premise

Everything hinges on a single vision-language model: the same model decides which objects are real versus hallucinated, where they sit in 3D, how they change over time, and which camera fixes are needed, with no learned geometric verification or 3D ground truth to check its physical predictions.

Editorial extensions

If this is right

  • Edits become local: expanding a scene, changing a style, or substituting a destroyed asset modifies only the affected object records, so the rest of the world keeps its identity and layout.
  • Camera work is verifiable before production: proposed trajectories are rendered and repaired against occlusion, framing, and collision, so director intent is checked against geometry rather than guessed.
  • The same world state serves multiple outputs, including video creation, shot planning, keyframes, and game-like interaction, without regenerating the scene for each use.
  • Video fidelity can be layered on: an off-the-shelf video model renders the geometry-grounded observations, so visual quality improves without altering the underlying world state.
  • Object identity persists across time and viewpoints, reducing the subject drift and background inconsistency typical of one-shot video synthesis.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper: StateFlow's transition plan could be driven by a learned dynamics model or physics simulator rather than a vision-language model, making state evolution testable against physical ground truth.
  • Beyond the paper: the object-centric state table could serve as an interchange format between generative models and game engines, since it already exposes geometry, pose, and semantic attributes as editable records.
  • Beyond the paper: a natural stress test is edit locality, measuring whether a one-line intent change alters only the intended object records and leaves unrelated scene elements bit-identical, which the paper reports qualitatively but does not quantify.
  • Beyond the paper: the dual-view construction could be extended to additional views to see whether conflict resolution scales beyond front-and-bird's-eye views or whether error rates grow with more views.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 5 minor

Summary. The paper proposes StateFlow, a state-centric framework for generative previsualization. Instead of one-shot image or video synthesis, StateFlow maintains a persistent object-centric 3D world state W_t = {o_i^t = (g_i^t, p_i^t, s_i^t)} with three stages: construction (prior-guided conflict-aware dual-view initialization), evolution (intent-guided structured state transitions), and access (world-state camera planning with render-feedback reflection). The framework is instantiated with off-the-shelf components: Gemini 3.1 for VLM reasoning, Nano Banana 2 for image generation, Hunyuan3D for lifting crops to 3D assets, and Seedance2 for video enhancement. Quantitative comparisons are reported against video-generation baselines (VBench, Table 1), scene-generation baselines (CLIP-I/T, HPS, Q-Align, Table 2), and a combined user study/MLLM evaluation (Table 3), along with ablations of construction and camera-planning components.

Significance. If the central claim is fully established, StateFlow is a meaningful step toward controllable previsualization: the formulation cleanly separates world state, evolution, and access; the construction and camera-planning stages are training-free and model-agnostic; and the paper includes a human study alongside automatic metrics. The manuscript is honest about its dependence on third-party models and explicitly frames the world state as an editable representation. However, the evidence is incomplete: the 'evolving' pillar of the claim has no quantitative evaluation, the main automatic evaluator is the same model family that performs planning, and the video-level comparison may conflate StateFlow's contribution with the Seedance enhancement backbone. Because these gaps bear directly on two of the three named contributions, the current evidence is not sufficient for acceptance, though the deficiencies appear fixable within the manuscript's scope.

major comments (5)
  1. [§4.5 and §3.4 (Eq. 6)] The State Evolution ablation is asserted but not measured. Section 4.5 states that replacing selective state updates with direct regeneration 'weakens identity preservation and spatial consistency across edits,' and that results are reported alongside baselines in Table 3, but no row or column in Table 3 corresponds to this ablation, no metric is given, and no figure is referenced. Since 'evolving' is one of the paper's three core contributions, the claim that structured state transitions outperform full-scene regeneration for iterative editing is currently unsupported. The authors should add quantitative experiments on multi-turn edits (e.g., identity preservation, spatial consistency, layout plausibility before and after edits, with and without selective updates), plus a user study on edit sequences, not only single-turn outputs.
  2. [§4.1, §4.3, Table 1] The video-level comparison is confounded by the Seedance2 enhancement. Implementation details state that Seedance2 is used as the video generation backbone for downstream outputs, while Table 1 lists Seedance2.0 as a baseline. It is therefore unclear whether the 'Ours' videos evaluated in Table 1, Table 3 (video-level), and Fig. 4 are pure geometry-grounded renders or Seedance2-enhanced videos. If the latter, the VBench scores for 'Ours' partly reflect the enhancement model's capabilities rather than StateFlow's state representation, making the comparison with the Seedance2.0 baseline unfair. The paper should report the rendered-only and enhanced results separately, or apply the same enhancement to all baselines, and state explicitly which output is used in each evaluation.
  3. [§4.4 and §4.1] The automatic MLLM evaluation uses Gemini 3.1, the same model family that performs scene planning, state transition prediction, and camera reflection. This creates a self-preference risk: the evaluator may systematically favor layouts and trajectories it proposed. The human study partially offsets this, but the paper reports both scores side by side as if they were equally independent, and does not report human-MLLM agreement or any analysis of divergence. The authors should either use an independent evaluator, report the MLLM scores as a secondary diagnostic with an explicit caveat, or provide evidence (e.g., per-prompt correlation) that the MLLM scores align with human judgments.
  4. [§3.3 (Eq. 4) and §3.5 (Eq. 9)] The construction and camera-planning objectives are not specified in enough detail to be reproduced or audited. Equation (4) contains four loss terms with weights lambda_b, lambda_v, lambda_p that are never given, and the text does not describe the optimization algorithm, the box parameterization, the number of iterations, or how the front-view identity term is computed. Similarly, the trajectory selection score J in Eq. (9) is described only qualitatively. Given that these objectives are central to the method's claimed generality, the authors should provide the full objective definitions, weights, solver details, and, ideally, code or a detailed pseudo-code appendix.
  5. [Tables 2, 3] No significance tests, confidence intervals, or per-prompt variance are reported for any quantitative comparison. In the user study, N=30 participants rate 12 prompts per setting, but only mean scores are shown, so it is impossible to tell whether differences between StateFlow and the strongest baselines (e.g., SynCity in Table 2, Seedance2.0 in Table 3) are meaningful relative to inter-participant or inter-prompt variance. The authors should add standard deviations or CIs and, where feasible, paired tests; otherwise the claimed overall superiority is not statistically grounded.
minor comments (5)
  1. [Fig. 2] The state table in Fig. 2 contains the typo 'Solar Pbanel' in the rightmost table; it should read 'Solar Panel'.
  2. [§4.4] The user study description says attention-check items were inserted but does not report how many participants failed them or whether their responses were excluded; this information is needed to judge data quality.
  3. [§4.3, Table 1] The VBench evaluation does not state the video resolution, duration, or frame count used for the comparison; since the paper compares methods that may generate different aspect ratios and lengths, these details should be reported for fairness.
  4. [§4.2, Table 2] PartCrafter is described as a part-aware 3D object generation method, while the comparison claims to be on scene generation; the task mismatch should be acknowledged or the baseline matched to the scene-generation setting.
  5. [§4.4] The evaluation prompts used for the MLLM judge are not given; to let readers gauge possible bias and to support replication, the full evaluation prompt should be included in the appendix.

Circularity Check

1 steps flagged · score 2.0 of 10

Equations are self-contained; the only real circularity-adjacent issue is the same VLM serving as both planner and MLLM evaluator.

  1. other [Sec. 4.1 Implementation Details and Sec. 4.4 MLLM-based Evaluation and User Study, Table 3]
    "We use Gemini 3.1 [9] as the vision-language model for scene understanding, intent parsing, layout reasoning, state transition planning, and camera feedback reflection. ... In addition, we use gemini-3.1 as an MLLM evaluation to provide complementary automatic assessment."

    The automatic evaluator is the same Gemini 3.1 model that generated the front/BEV priors, predicted grounding and transitions, and proposed/refined camera trajectories. Thus the MLLM columns of Table 3 score each output for agreement with the very model's priors that produced it, a self-preference loop rather than independent measurement. This is not an equation-level derivation identity: no W_t or score is defined in terms of the other, and the same table reports a 30-participant user study while Tables 1-2 use VBench, CLIP, HPS, and Q-Align, so the central claim does not reduce to the confounded automatic scores.

full rationale

Walking the derivation chain: Eqs. (1)-(2) define W_t and the build/evolve/access roles; Eqs. (3)-(5) construct W_0 from generated front/BEV views and VLM priors; Eq. (6) updates the state table; Eqs. (7)-(9) propose and repair camera trajectories. None of these equations defines a predicted quantity as the same expression as an input; the construction loss in Eq. (4) optimizes box parameters against front/BEV/VLM/physical terms, and the camera selection in Eq. (9) minimizes a separate feasibility score over repair candidates. The central comparison uses external metrics (VBench, CLIP-I/CLIP-T, HPS V2, Q-Align) and a 30-participant human study, so it is not a self-citation chain. The one genuine overlap is evaluative: Gemini 3.1 is both the internal planner and the automatic MLLM judge in Table 3, which can bias the automatic half of those cells. That confound is partially offset by the human-study half and external benchmarks. Separately, the claimed 'State Evolution' benefit is supported only by one sentence in Sec. 4.5 without metrics; this is a missing-evidence concern, not a circular derivation, because nothing about Eq. (6) forces the stated identity-preservation advantage. Overall no load-bearing circularity; score 2 for the minor self-evaluation confound.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The framework rests on assumptions about VLM reliability and generated-image coherence rather than on a mathematical derivation. All listed axioms are domain assumptions with no independent validation in the paper.

free parameters (3)
  • Eq. (4) loss weights lambda_b, lambda_v, lambda_p = not reported
    Balance BEV layout, VLM prior, and physical plausibility terms in the box-refinement objective; values are not given in the text.
  • Camera reflection search size M and maximum rounds = not reported
    Eq. (9) samples M local repair candidates around the current trajectory; the iteration cap is also unspecified, affecting trajectory quality and runtime.
  • Category size prior s_bar_i for 3D boxes = not reported
    Used in the grounding-aware lifting operator Eq. (3), sourced from VLM or category priors with no reported uncertainty.
assumptions (4)
  • domain assumption Independently generated front-view and BEV images contain sufficient and complementarily consistent information to reconstruct a coherent 3D world.
    State construction Sec. 3.3 relies on these two views as the only geometric inputs; no 3D ground truth is used.
  • domain assumption A VLM's semantic and physical priors can correctly resolve cross-view object-count and spatial conflicts.
    Conflict resolution and grounding prediction in Sec. 3.3 are delegated to Gemini 3.1 without error analysis.
  • domain assumption Image-to-3D lifting of object crops produces geometry that matches the optimized 3D boxes and scene layout.
    Sec. 3.3 generates each object's geometry independently with Hunyuan3D and does not check consistency against the jointly optimized boxes.
  • domain assumption Heuristic render-feedback checklists such as visibility, framing, collision, and motion capture trajectory quality.
    Sec. 3.5 uses these criteria in the cost J for camera refinement; there is no validation that the checklist aligns with human shot preferences.

how reviews work

0 comments
Cite this review

Pith. "Pith review of StateFlow: Building, Evolving, and Accessing 3D World States for Previsualization." pith.science (2026). https://pith.science/paper/QIEXYAXK

@misc{pith2026260812314,
  author       = {Pith},
  title        = {Pith review of: StateFlow: Building, Evolving, and Accessing 3D World States for Previsualization},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/QIEXYAXK}},
  note         = {Machine review of arXiv:2608.12314}
}
read the original abstract

Previsualization is an intermediate layer between ideas and production in film, games, architecture, and urban design. It lets creators iteratively refine scenes, actions, cameras, and spatial-temporal dynamics. Yet existing generative methods rely on simple prompts to jointly control all of these factors through one-shot image or video synthesis, offering weak controllability and limited support for iterative editing. Fundamentally, a world comprises multiple elements with geometry, appearance, and other attributes, together with cameras. Different frames are produced through local modifications or recombinations of this shared state, which is otherwise largely reused. Therefore, we argue that the missing component is an explicit and persistent working state. To address this, we present StateFlow, a state-centric framework for generative previsualization. Rather than generating videos in one shot, StateFlow uses an editable 3D world to organize scene structure, evolution, and cameras, while off-the-shelf video models enhance visual quality when higher fidelity is desired. This world is maintained as a persistent structured 3D state of scene elements and camera configurations, serving as the core working representation for previsualization. Built on this insight, StateFlow has three stages to construct, evolve, and access the world state. State construction lifts generated 2D content into a coherent 3D world through prior-guided, conflict-aware dual-view initialization, while State evolution translates user intent into structured state transitions while preserving world memory, avoiding full-scene regeneration for each edit. State access uses render-feedback reflection to refine camera plans into visually feasible trajectories, avoiding reliance on VLM semantics alone. Experiments show that StateFlow produces high-quality 3D worlds for video creation and game-like prototyping.

Figures

Figures reproduced from arXiv: 2608.12314 by the authors.

Figure 1
Figure 1. StateFlow turns a creative prompt into a complete, editable 3D world. Creators can construct scenes from reference images, [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. StateFlow turns creative intent into a persistent 3D world that can be constructed, evolved, and accessed for downstream creation. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Qualitative comparison on Scene Generation [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison on previsualization applications. Case 1 and Case 3 showcase [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 5
Figure 5. Figure 5: Visualization of Ablation Study side by side in anonymized and randomized order. Partic￾ipants rate each method on a 5-point Likert scale, where 1 indicates very poor quality and 5 indicates excellent quality. We also clarify that a score of 5 does not require perfecti…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

62 extracted references · 39 canonical work pages

  1. [1]

    A collaborative previsualiza- tion tool for filmmaking in virtual reality

    Dui Ardal, Simon Alexandersson, Mirko Lempert, and Andr´e Tiago Abelho Pereira. A collaborative previsualiza- tion tool for filmmaking in virtual reality. InProceedings of the 16th ACM SIGGRAPH European Conference on Visual Media Production, pages 1–10, 2019. 2, 3

    work page 2019

  2. [2]

    Stable video diffusion: Scaling latent video diffusion models to large datasets.arXiv preprint arXiv:2311.15127, 2023

    Andreas Blattmann, Tim Dockhorn, Sumith Kulal, Daniel Mendelevitch, Maciej Kilian, Dominik Lorenz, Yam Levi, Zion English, Vikram V oleti, Adam Letts, et al. Stable video diffusion: Scaling latent video diffusion models to large datasets.arXiv preprint arXiv:2311.15127, 2023. 3

    arXiv 2023

  3. [3]

    Video generation models as world simulators.OpenAI Blog, 1(8):1, 2024

    Tim Brooks, Bill Peebles, Connor Holmes, Will DePue, Yufei Guo, Leo Jing, David Schnurr, Joe Taylor, Troy Luh- man, Eric Luhman, et al. Video generation models as world simulators.OpenAI Blog, 1(8):1, 2024. 2

    work page 2024

  4. [4]

    Ge- nie: Generative interactive environments

    Jake Bruce, Michael D Dennis, Ashley Edwards, Jack Parker-Holder, Yuge Shi, Edward Hughes, Matthew Lai, Aditi Mavalankar, Richie Steigerwald, Chris Apps, et al. Ge- nie: Generative interactive environments. InForty-first Inter- national Conference on Machine Learning, 2024. 2

    2024

  5. [5]

    Camera control in computer graphics

    Marc Christie, Patrick Olivier, and Jean-Marie Normand. Camera control in computer graphics. InComputer graphics forum, pages 2197–2218. Wiley Online Library, 2008. 2

    work page 2008

  6. [6]

    Syncity: Training-free gen- eration of 3d worlds

    Paul Engstler, Aleksandar Shtedritski, Iro Laina, Christian Rupprecht, and Andrea Vedaldi. Syncity: Training-free gen- eration of 3d worlds. InProceedings of the IEEE/CVF In- ternational Conference on Computer Vision, pages 27585– 27595, 2025. 2, 3, 7

    work page 2025

  7. [7]

    Tape: Tangible augmented previz environment for filmmaking

    Guangzheng Fei and Dake Liu. Tape: Tangible augmented previz environment for filmmaking. InProceedings of the Eleventh International Symposium of Chinese CHI, pages 251–262, 2023. 3

    work page 2023

  8. [8]

    Layoutgpt: Compositional visual plan- ning and generation with large language models.Advances in Neural Information Processing Systems, 36:18225–18250,

    Weixi Feng, Wanrong Zhu, Tsu-jui Fu, Varun Jampani, Ar- jun Akula, Xuehai He, Sugato Basu, Xin Eric Wang, and William Yang Wang. Layoutgpt: Compositional visual plan- ning and generation with large language models.Advances in Neural Information Processing Systems, 36:18225–18250,

    work page

Show all 62 references
  1. [9]

    Gemini 3.1 Pro Preview.https://ai.google

    Google. Gemini 3.1 Pro Preview.https://ai.google. dev/gemini- api/docs/models/gemini- 3.1- pro-preview, 2026. Accessed: 2026-05-12. 7

    2026

  2. [10]

    Nano Banana 2: Google’s latest AI image genera- tion model.https://blog.google/innovation- and-ai/technology/ai/nano-banana-2/, 2026

    Google. Nano Banana 2: Google’s latest AI image genera- tion model.https://blog.google/innovation- and-ai/technology/ai/nano-banana-2/, 2026. Accessed: 2026-05-12. 7

    2026

  3. [11]

    Artiscene: Language-driven artistic 3d scene generation through image intermediary

    Zeqi Gu, Yin Cui, Zhaoshuo Li, Fangyin Wei, Yunhao Ge, Jinwei Gu, Ming-Yu Liu, Abe Davis, and Yifan Ding. Artiscene: Language-driven artistic 3d scene generation through image intermediary. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 2891– 29...

    2025

  4. [12]

    From camera-eye to ai: Exploring the interplay of cinematography and compu- tational visual storytelling

    Brett A Halperin and Stephanie M Lukin. From camera-eye to ai: Exploring the interplay of cinematography and compu- tational visual storytelling. InProceedings of the 2025 CHI Conference on Human Factors in Computing Systems, pages 1–18, 2025. 2

    2025

  5. [13]

    Toonflow-app: Open-source ai tool for animated short drama creation.https://github.com/HBAI- Ltd/Toonflow-app, 2026

    HBAI-Ltd. Toonflow-app: Open-source ai tool for animated short drama creation.https://github.com/HBAI- Ltd/Toonflow-app, 2026. GitHub repository, accessed 2026-08-08. 4

    2026

  6. [14]

    Videoclaw: Ai creative video produc- tion agent.https : / / github

    HITsz-TMG. Videoclaw: Ai creative video produc- tion agent.https : / / github. com / HITsz - TMG / VideoClaw, 2026. GitHub repository, accessed 2026-08-

    2026

  7. [15]

    Vimax: Agentic video gen- eration.arXiv preprint arXiv:2606.07649, 2026

    Lingxuan Huang, Sizhe He, Hengji Zhou, Liqiang Nie, Lianghao Xia, and Chao Huang. Vimax: Agentic video gen- eration.arXiv preprint arXiv:2606.07649, 2026. 4

    2026 arXiv

  8. [16]

    Vbench: Comprehensive bench- mark suite for video generative models

    Ziqi Huang, Yinan He, Jiashuo Yu, Fan Zhang, Chenyang Si, Yuming Jiang, Yuanhan Zhang, Tianxing Wu, Qingyang Jin, Nattapol Chanpaisit, et al. Vbench: Comprehensive bench- mark suite for video generative models. InProceedings of the IEEE/CVF Conference on Computer Vision and Pa...

    2024

  9. [17]

    Midi: Multi-instance diffusion for single image to 3d scene generation

    Zehuan Huang, Yuan-Chen Guo, Xingqiao An, Yunhan Yang, Yangguang Li, Zi-Xin Zou, Ding Liang, Xihui Liu, Yan-Pei Cao, and Lu Sheng. Midi: Multi-instance diffusion for single image to 3d scene generation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Re...

    2025

  10. [18]

    One week in the future: Previs design futuring for hci research

    Alexander Ivanov, Tim Au Yeung, Kathryn Blair, Kurtis Danyluk, Georgina Freeman, Marcus Friedel, Carmen Hull, Michael Yuk-Shing Hung, Sydney Pratte, and Wesley Wil- lett. One week in the future: Previs design futuring for hci research. InProceedings of the 2022 CHI Conference ...

    2022

  11. [19]

    Collagevis: Rapid previsualization tool for indie filmmaking using video collages

    Hye-Young Jo, Ryo Suzuki, and Yoonji Kim. Collagevis: Rapid previsualization tool for indie filmmaking using video collages. InProceedings of the 2024 CHI Conference on Human Factors in Computing Systems, pages 1–16, 2024. 3

    2024

  12. [20]

    Hunyuanvideo: A systematic framework for large video generative models.arXiv preprint arXiv:2412.03603, 2024

    Weijie Kong, Qi Tian, Zijian Zhang, Rox Min, Zuozhuo Dai, Jin Zhou, Jiangfeng Xiong, Xin Li, Bo Wu, Jianwei Zhang, et al. Hunyuanvideo: A systematic framework for large video generative models.arXiv preprint arXiv:2412.03603, 2024. 3

    2024 arXiv

  13. [21]

    Martian world model: Con- trollable video synthesis with physically accurate 3d recon- structions.Advances in Neural Information Processing Sys- tems, 38, 2026

    Longfei Li, Zhiwen Fan, Wenyan Cong, Xinhang Liu, Yuyang Yin, Matt Foutter, Panwang Pan, Chenyu You, Yue Wang, Zhangyang Wang, et al. Martian world model: Con- trollable video synthesis with physically accurate 3d recon- structions.Advances in Neural Information Processing Sys...

    2026

  14. [22]

    Diffusion4d: Fast spatial-temporal consis- tent 4d generation via video diffusion models.arXiv preprint arXiv:2405.16645, 2024

    Hanwen Liang, Yuyang Yin, Dejia Xu, Hanxue Liang, Zhangyang Wang, Konstantinos N Plataniotis, Yao Zhao, and Yunchao Wei. Diffusion4d: Fast spatial-temporal consis- tent 4d generation via video diffusion models.arXiv preprint arXiv:2405.16645, 2024

    2024 arXiv

  15. [23]

    Partcrafter: Structured 3d mesh generation via compo- sitional latent diffusion transformers.arXiv preprint arXiv:2506.05573, 2025

    Yuchen Lin, Chenguo Lin, Panwang Pan, Honglei Yan, Yiqiang Feng, Yadong Mu, and Katerina Fragkiadaki. Partcrafter: Structured 3d mesh generation via compo- sitional latent diffusion transformers.arXiv preprint arXiv:2506.05573, 2025. 3, 7

    2025 arXiv

  16. [24]

    The director’s lens: an intelligent assis- tant for virtual cinematography

    Christophe Lino, Marc Christie, Roberto Ranon, and William Bares. The director’s lens: an intelligent assis- tant for virtual cinematography. InProceedings of the 19th ACM international conference on Multimedia, pages 323– 332, 2011. 2

    2011

  17. [25]

    Chat- cam: Empowering camera control through conversational ai.Advances in Neural Information Processing Systems, 37: 54483–54506, 2024

    Xinhang Liu, Yu-Wing Tai, and Chi-Keung Tang. Chat- cam: Empowering camera control through conversational ai.Advances in Neural Information Processing Systems, 37: 54483–54506, 2024. 6

    2024

  18. [26]

    Experiments in the use of game technology for pre-visualization

    Michael Nitsche. Experiments in the use of game technology for pre-visualization. InProceedings of the 2008 Confer- ence on Future Play: Research, Play, Share, pages 160–165,

    2008

  19. [27]

    A collab- orative real time previsualization tool for video games and film

    Lesley Northam, Joe Istead, and Craig S Kaplan. A collab- orative real time previsualization tool for video games and film. InACM SIGGRAPH 2012 Posters, pages 1–1. 2012. 3

    2012

  20. [28]

    Learning transferable visual models from natural language supervi- sion

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervi- sion. InInternational conference on machine learning, p...

    2021

  21. [29]

    Infinite photore- alistic worlds using procedural generation

    Alexander Raistrick, Lahav Lipson, Zeyu Ma, Lingjie Mei, Mingzhe Wang, Yiming Zuo, Karhan Kayan, Hongyu Wen, Beining Han, Yihan Wang, Alejandro Newell, Hei Law, Ankit Goyal, Kaiyu Yang, and Jia Deng. Infinite photore- alistic worlds using procedural generation. InProceedings o...

    2023

  22. [30]

    Infinigen indoors: Photorealistic in- door scenes using procedural generation

    Alexander Raistrick, Lingjie Mei, Karhan Kayan, David Yan, Yiming Zuo, Beining Han, Hongyu Wen, Meenal Parakh, Stamatis Alexandropoulos, Lahav Lipson, Zeyu Ma, and Jia Deng. Infinigen indoors: Photorealistic in- door scenes using procedural generation. InProceedings of the IEE...

    2024

  23. [31]

    Dreamgaussian4d: Genera- tive 4d gaussian splatting.arXiv preprint arXiv:2312.17142,

    Jiawei Ren, Liang Pan, Jiaxiang Tang, Chi Zhang, Ang Cao, Gang Zeng, and Ziwei Liu. Dreamgaussian4d: Genera- tive 4d gaussian splatting.arXiv preprint arXiv:2312.17142,

    arXiv

  24. [32]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj ¨orn Ommer. High-resolution image synthesis with latent diffusion models. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 10684–10695, 2022. 2

    2022

  25. [33]

    Seedance 2.0: Advanc- ing video generation for world complexity.arXiv preprint arXiv:2604.14148, 2026

    Team Seedance, De Chen, Liyang Chen, Xin Chen, Ying Chen, Zhuo Chen, Zhuowei Chen, Feng Cheng, Tianheng Cheng, Yufeng Cheng, et al. Seedance 2.0: Advanc- ing video generation for world complexity.arXiv preprint arXiv:2604.14148, 2026. 2, 3, 7, 9

    2026 arXiv

  26. [34]

    Animaker: Multi-agent animated storytelling with mcts-driven clip generation

    Haoyuan Shi, Yunxin Li, Xinyu Chen, Longyue Wang, Bao- tian Hu, and Min Zhang. Animaker: Multi-agent animated storytelling with mcts-driven clip generation. InProceed- ings of the SIGGRAPH Asia 2025 Conference Papers, pages 1–11, 2025. 3, 9 12

    2025

  27. [35]

    Scenemaker: Open-set 3d scene generation with decoupled de-occlusion and pose es- timation model.arXiv preprint arXiv:2512.10957, 2025

    Yukai Shi, Weiyu Li, Zihao Wang, Hongyang Li, Xingyu Chen, Ping Tan, and Lei Zhang. Scenemaker: Open-set 3d scene generation with decoupled de-occlusion and pose es- timation model.arXiv preprint arXiv:2512.10957, 2025. 2, 3

    2025

  28. [36]

    Layoutvlm: Differentiable optimization of 3d layout via vision-language models

    Fan-Yun Sun, Weiyu Liu, Siyi Gu, Dylan Lim, Goutam Bhat, Federico Tombari, Manling Li, Nick Haber, and Jia- jun Wu. Layoutvlm: Differentiable optimization of 3d layout via vision-language models. InProceedings of the Computer Vision and Pattern Recognition Conference, pages 29...

    2025

  29. [37]

    Dimensionx: Create any 3d and 4d scenes from a single image with controllable video diffusion.arXiv preprint arXiv:2411.04928, 2024

    Wenqiang Sun, Shuo Chen, Fangfu Liu, Zilong Chen, Yueqi Duan, Jun Zhang, and Yikai Wang. Dimensionx: Create any 3d and 4d scenes from a single image with controllable video diffusion.arXiv preprint arXiv:2411.04928, 2024. 3

    2024 arXiv

  30. [38]

    Worldplay: Towards long-term geometric consistency for real-time interactive world modeling.arXiv preprint arXiv:2512.14614, 2025

    Wenqiang Sun, Haiyu Zhang, Haoyuan Wang, Junta Wu, Zehan Wang, Zhenwei Wang, Yunhong Wang, Jun Zhang, Tengfei Wang, and Chunchao Guo. Worldplay: Towards long-term geometric consistency for real-time interactive world modeling.arXiv preprint arXiv:2512.14614, 2025

    2025 arXiv

  31. [39]

    Dreamgaussian: Generative gaussian splatting for ef- ficient 3d content creation

    Jiaxiang Tang, Jiawei Ren, Hang Zhou, Ziwei Liu, and Gang Zeng. Dreamgaussian: Generative gaussian splatting for ef- ficient 3d content creation. InInternational Conference on Learning Representations, pages 33879–33896, 2024

    2024

  32. [40]

    Hy-world 2.0: A multi-modal world model for reconstructing, generating, and simulating 3d worlds

    HY Team, Chenjie Cao, Xuhui Zuo, Zhenwei Wang, Yisu Zhang, Junta Wu, Zhenyang Liu, Yuning Gong, Yang Liu, Bo Yuan, et al. Hy-world 2.0: A multi-modal world model for reconstructing, generating, and simulating 3d worlds. arXiv e-prints, pages arXiv–2604, 2026

    2026

  33. [41]

    Sam 3d: 3dfy anything in images

    SAM 3D Team, Xingyu Chen, Fu-Jen Chu, Pierre Gleize, Kevin J Liang, Alexander Sax, Hao Tang, Weiyao Wang, Michelle Guo, Thibaut Hardin, Xiang Li, Aohan Lin, Jiawei Liu, Ziqi Ma, Anushka Sagar, Bowen Song, Xiaodong Wang, Jianing Yang, Bowen Zhang, Piotr Doll´ar, Georgia Gkioxar...

    2025

  34. [42]

    Hunyuan3d 2.5: Towards high- fidelity 3d assets generation with ultimate details, 2025

    Tencent Hunyuan3D Team. Hunyuan3d 2.5: Towards high- fidelity 3d assets generation with ultimate details, 2025. 7

    2025

  35. [43]

    Wan: Open and advanced large-scale video gen- erative models.arXiv preprint arXiv:2503.20314, 2025

    Team Wan, Ang Wang, Baole Ai, Bin Wen, Chaojie Mao, Chen-Wei Xie, Di Chen, Feiwu Yu, Haiming Zhao, Jianxiao Yang, et al. Wan: Open and advanced large-scale video gen- erative models.arXiv preprint arXiv:2503.20314, 2025. 2, 3, 9

    2025 arXiv

  36. [44]

    Worldgen: From text to traversable and interactive 3d worlds.arXiv preprint arXiv:2511.16825, 2025

    Dilin Wang, Hyunyoung Jung, Tom Monnier, Kihyuk Sohn, Chuhang Zou, Xiaoyu Xiang, Yu-Ying Yeh, Di Liu, Zix- uan Huang, Thu Nguyen-Phuoc, et al. Worldgen: From text to traversable and interactive 3d worlds.arXiv preprint arXiv:2511.16825, 2025. 2, 3

    2025

  37. [45]

    Q-align: Teaching lmms for visual scoring via discrete text-defined levels.arXiv preprint arXiv:2312.17090, 2023

    Haoning Wu, Zicheng Zhang, Weixia Zhang, Chaofeng Chen, Liang Liao, Chunyi Li, Yixuan Gao, Annan Wang, Erli Zhang, Wenxiu Sun, et al. Q-align: Teaching lmms for visual scoring via discrete text-defined levels.arXiv preprint arXiv:2312.17090, 2023. 7

    2023 arXiv

  38. [46]

    Auto- mated movie generation via multi-agent cot planning.arXiv preprint arXiv:2503.07314, 2025

    Weijia Wu, Zeyu Zhu, and Mike Zheng Shou. Auto- mated movie generation via multi-agent cot planning.arXiv preprint arXiv:2503.07314, 2025. 3, 9

    2025 arXiv

  39. [47]

    Human preference score: Better aligning text- to-image models with human preference

    Xiaoshi Wu, Keqiang Sun, Feng Zhu, Rui Zhao, and Hong- sheng Li. Human preference score: Better aligning text- to-image models with human preference. InProceedings of the IEEE/CVF International Conference on Computer Vi- sion, pages 2096–2105, 2023. 7

    2023

  40. [48]

    Structured 3d latents for scalable and versatile 3d generation

    Jianfeng Xiang, Zelong Lv, Sicheng Xu, Yu Deng, Ruicheng Wang, Bowen Zhang, Dong Chen, Xin Tong, and Jiaolong Yang. Structured 3d latents for scalable and versatile 3d generation. In2025 IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition (CVPR), pages 21469–21480...

    2025

  41. [49]

    Tip4gen: Text to immersive panorama 4d scene generation

    Ke Xing, Hanwen Liang, Dejia Xu, Yuyang Yin, Konstanti- nos N Plataniotis, Yao Zhao, and Yunchao Wei. Tip4gen: Text to immersive panorama 4d scene generation. InPro- ceedings of the 33rd ACM International Conference on Mul- timedia, pages 9267–9276, 2025. 3

    2025

  42. [50]

    Llplace: The 3d in- door scene layout generation and editing via large language model.arXiv preprint arXiv:2406.03866, 2024

    Yixuan Yang, Junru Lu, Zixiang Zhao, Zhen Luo, James JQ Yu, Victor Sanchez, and Feng Zheng. Llplace: The 3d in- door scene layout generation and editing via large language model.arXiv preprint arXiv:2406.03866, 2024. 3

    2024 arXiv

  43. [51]

    Holodeck: Language guided gen- eration of 3d embodied ai environments

    Yue Yang, Fan-Yun Sun, Luca Weihs, Eli VanderBilt, Al- varo Herrasti, Winson Han, Jiajun Wu, Nick Haber, Ranjay Krishna, Lingjie Liu, et al. Holodeck: Language guided gen- eration of 3d embodied ai environments. InProceedings of the IEEE/CVF Conference on Computer Vision and P...

    2024

  44. [52]

    Cogvideox: Text-to-video diffusion models with an expert transformer.arXiv preprint arXiv:2408.06072, 2024

    Zhuoyi Yang, Jiayan Teng, Wendi Zheng, Ming Ding, Shiyu Huang, Jiazheng Xu, Yuanming Yang, Wenyi Hong, Xiao- han Zhang, Guanyu Feng, et al. Cogvideox: Text-to-video diffusion models with an expert transformer.arXiv preprint arXiv:2408.06072, 2024. 2, 3

    2024 arXiv

  45. [53]

    Matrix-3d: Omnidirectional explorable 3d world generation.arXiv preprint arXiv:2508.08086, 2025

    Zhongqi Yang, Wenhang Ge, Yuqi Li, Jiaqi Chen, Haoyuan Li, Mengyin An, Fei Kang, Hua Xue, Baixin Xu, Yuyang Yin, et al. Matrix-3d: Omnidirectional explorable 3d world generation.arXiv preprint arXiv:2508.08086, 2025. 3

    2025 arXiv

  46. [54]

    Mmgdreamer: Mixed-modality graph for geometry-controllable 3d indoor scene generation

    Zhifei Yang, Keyang Lu, Chao Zhang, Jiaxing Qi, Hanqi Jiang, Ruifei Ma, Shenglin Yin, Yifan Xu, Mingzhe Xing, Zhen Xiao, et al. Mmgdreamer: Mixed-modality graph for geometry-controllable 3d indoor scene generation. InPro- ceedings of the AAAI Conference on Artificial Intellige...

    2025

  47. [55]

    Flowscene: Style-consistent indoor scene genera- tion with multimodal graph rectified flow.arXiv preprint arXiv:2603.19598, 2026

    Zhifei Yang, Guangyao Zhai, Keyang Lu, YuYang Yin, Chao Zhang, Zhen Xiao, Jieyi Long, Nassir Navab, and Yikai Wang. Flowscene: Style-consistent indoor scene genera- tion with multimodal graph rectified flow.arXiv preprint arXiv:2603.19598, 2026. 3

    2026

  48. [56]

    4dgen: Grounded 4d content gen- eration with spatial-temporal consistency.arXiv preprint arXiv:2312.17225, 2023

    Yuyang Yin, Dejia Xu, Zhangyang Wang, Yao Zhao, and Yunchao Wei. 4dgen: Grounded 4d content gen- eration with spatial-temporal consistency.arXiv preprint arXiv:2312.17225, 2023. 3

    2023 arXiv

  49. [57]

    Panoworld-x: Generating explorable panoramic worlds via sphere-aware video diffusion.arXiv preprint arXiv:2509.24997, 2025

    Yuyang Yin, HaoXiang Guo, Fangfu Liu, Mengyu Wang, Hanwen Liang, Eric Li, Yikai Wang, Xiaojie Jin, Yao Zhao, and Yunchao Wei. Panoworld-x: Generating explorable panoramic worlds via sphere-aware video diffusion.arXiv preprint arXiv:2509.24997, 2025. 3

    2025

  50. [58]

    A system- atic review of architectural design collaboration in immersive virtual environments.Designs, 6(5):93, 2022

    Rongrong Yu, Ning Gu, Gun Lee, and Ayaz Khan. A system- atic review of architectural design collaboration in immersive virtual environments.Designs, 6(5):93, 2022. 3

    2022

  51. [59]

    Anime: Adaptive multi-agent planning for long animation generation

    Lisai Zhang, Baohan Xu, Siqian Yang, Mingyu Yin, Jing Liu, Chao Xu, Siqi Wang, Yidi Wu, Yuxin Hong, Zihao 13 Zhang, et al. Anime: Adaptive multi-agent planning for long animation generation. InProceedings of the SIGGRAPH Asia 2025 Posters, pages 1–3. 2025. 3

    2025

  52. [60]

    Worldstereo: Bridging camera-guided video generation and scene re- construction via 3d geometric memories.arXiv preprint arXiv:2603.02049, 2026

    Yisu Zhang, Chenjie Cao, Tengfei Wang, Xuhui Zuo, Junta Wu, Jianke Zhu, and Chunchao Guo. Worldstereo: Bridging camera-guided video generation and scene re- construction via 3d geometric memories.arXiv preprint arXiv:2603.02049, 2026. 3

    2026

  53. [61]

    Hunyuan3d 2.0: Scaling diffu- sion models for high resolution textured 3d assets generation

    Zibo Zhao, Zeqiang Lai, Qingxiang Lin, Yunfei Zhao, Haolin Liu, Shuhui Yang, Yifei Feng, Mingxin Yang, Sheng Zhang, Xianghui Yang, et al. Hunyuan3d 2.0: Scaling diffu- sion models for high resolution textured 3d assets generation. arXiv preprint arXiv:2501.12202, 2025. 3

    2025 arXiv

  54. [62]

    Imaginarium: Vision-guided high- quality 3d scene layout generation.ACM Transactions on Graphics (TOG), 44(6):1–24, 2025

    Xiaoming Zhu, Xu Huang, Qinghongbing Xie, Zhi Deng, Junsheng Yu, Yirui Guan, Zhongyuan Liu, Lin Zhu, Qijun Zhao, Ligang Liu, et al. Imaginarium: Vision-guided high- quality 3d scene layout generation.ACM Transactions on Graphics (TOG), 44(6):1–24, 2025. 3 14

    2025

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.