Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

HairWeaver animates a single photo with physically plausible hair motion by learning control from simulation and discarding the style adapter at inference.

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 ↗

T0 review · deepseek-v4-flash

2026-08-03 00:11 UTC pith:264DBNRV

load-bearing objection The temporary domain-LoRA trick is genuinely clever, but the evaluation doesn't yet back the state-of-the-art claim, especially for real images. the 3 major comments →

arxiv 2602.11117 v2 pith:264DBNRV submitted 2026-02-11 cs.CV

HairWeaver: Few-Shot Photorealistic Hair Motion Synthesis with Sim-to-Real Guided Video Diffusion

Di Chang , Ji Hou , Aljaz Bozic , Assaf Neuberger , Felix Juefei-Xu , Olivier Maury , Gene Wei-Chin Lin , Tuur Stuyck
show 3 more authors
Doug Roble Mohammad Soleymani Stephane Grabli
This is my paper
classification cs.CV
keywords hair motion synthesisimage-to-video animationvideo diffusion modelLoRA adaptationsim-to-real transferphysics-based simulationsingle-image animationsecondary dynamics
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 paper takes on a gap in single-image human animation: existing models move the body but leave hair stiff, because they treat hair as a static texture. HairWeaver claims that fine-grained hair control can be learned from a small synthetic dataset of physically simulated hair and then transferred to a photorealistic video diffusion backbone without carrying over the synthetic look. The transfer rests on a two-stage LoRA recipe: first adapt the backbone to the simulated domain, freeze that adapter, train a motion adapter on top of it, then remove the domain adapter at inference. If the recipe works, animators get hair that responds naturally to head and body motion while preserving the original photo's identity. The paper supports the claim with quantitative comparisons on a synthetic hair test set and a real-capture benchmark, plus a user study.

Core claim

The paper's central claim is that a temporary Sim2Real-Domain-LoRA acts as a bridge: it creates a stable, simulation-adapted feature space in which the Motion-Context-LoRA learns the mapping from hair and body motion conditions (a UVW strand-position buffer and body normal maps) to video frames. Once that mapping is learned, the bridge can be removed, and inference runs on the original photorealistic backbone plus the Motion-Context-LoRA alone. In the paper's notation, inference predicts noise with epsilon of theta and phi_P given x_t, t, c (Eq. 9), with the domain adapter gone. The result is claimed to be substantially better hair motion than pose-conditioned baselines on both the simulated

What carries the argument

The core mechanism is the two-stage train-then-discard LoRA protocol. Motion-Context-LoRA is a lightweight adapter that injects two motion signals into a frozen video diffusion transformer: body pose tokens are added directly to the noisy latent, and hair UVW-condition tokens are concatenated into the attention context. Sim2Real-Domain-LoRA is a separate adapter pretrained on synthetic videos, frozen during motion training, and discarded at inference. LoRA here means a low-rank weight adapter on the attention layers, so the backbone's original weights stay untouched.

Load-bearing premise

The load-bearing premise is that the motion mapping learned by Motion-Context-LoRA inside the Sim2Real-Domain-LoRA's adapted feature space stays valid after that adapter is deleted at inference; the paper does not directly compare this train-then-discard configuration against a version trained without the domain adapter.

What would settle it

Train the motion adapter in two configurations: with and without the Sim2Real-Domain-LoRA ever present, holding data and compute fixed, then compare hair-motion fidelity (for example cd-FVD on a synthetic test set with known ground-truth hair). If the never-used-domain-adapter model matches or beats the proposed discard version, the central transfer claim is falsified.

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

If this is right

  • If the transfer holds, single-image animation gains explicit, controllable hair motion without retraining a large video backbone.
  • The training need is modest: around a thousand synthetic videos with paired dense conditions, which is few-shot relative to the backbone's scale.
  • On the reported benchmarks, HairWeaver beats pose-only baselines on reconstruction and video-distribution metrics, suggesting that hair-specific conditioning is the missing ingredient for realistic secondary motion.
  • Because the domain adapter is discarded at inference, the method claims to preserve the base model's photorealism rather than inheriting the simulator's style.
  • The pose and hair encoders are lightweight, so the method is faster at inference than several of the compared animation systems.

Where Pith is reading between the lines

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

  • A direct test the paper does not run: train Motion-Context-LoRA without ever adding Sim2Real-Domain-LoRA, then compare motion fidelity against the train-then-discard version; if the simpler training matches it, the two-stage scheme is not load-bearing.
  • The recipe could generalize to other secondary dynamics (clothing, fur, fluids) wherever paired CG simulation with dense per-pixel conditions is available, since nothing in the protocol is hair-specific.
  • The use of a UVW strand buffer rather than sparse keypoints suggests that dense strand-space conditioning is what enables fine control; a comparison with optical-flow or direction-map conditions on the same backbone would isolate that factor.
  • On the real-capture benchmark the paper substitutes alpha maps and dense body pose for the UVW/normal conditions, so the real-video gains may come partly from the training data mix rather than the exact conditioning representation.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

3 major / 5 minor

Summary. HairWeaver proposes a diffusion-based pipeline for animating a single human image with controllable, physically plausible hair motion. The method generates a synthetic CG dataset with simulated hair strands and uses a two-stage training strategy: first a Sim2Real-Domain-LoRA adapts a frozen video DiT to the CG domain; then a Motion-Context-LoRA is trained to map body-pose and hair-UVW conditions to video, while the domain LoRA is frozen. At inference, the domain LoRA is discarded, leaving only the Motion-Context-LoRA to guide the original photorealistic backbone. The paper reports quantitative gains over baselines on a self-collected CG test set and on the NeRSemble dataset, plus a user study favoring HairWeaver.

Significance. If the central mechanism works as claimed, HairWeaver offers a practical recipe for transferring fine-grained, physics-based motion control from synthetic data to a frozen, photorealistic video foundation model using only lightweight adapters. The idea of training a motion adapter in a temporarily domain-adapted feature space and then discarding the domain adapter at inference is interesting and potentially useful beyond hair synthesis. The paper also provides a clear dataset-generation pipeline and a thorough set of standard metrics. However, the evaluation as presented does not yet support the state-of-the-art claim because the key inference-time design (train-with-then-discard) is not directly ablated on photorealistic data, and the photorealistic benchmark changes multiple variables at once.

major comments (3)
  1. [§3.4 (Eqs. 8–9); Table 3; Fig. 5] The central design is training the Motion-Context-LoRA ϕP in the frozen feature space of the Sim2Real-Domain-LoRA ϕD and then discarding ϕD at inference (Eq. 8 vs Eq. 9). This is never directly ablated on a photorealistic test set. Table 3's 'w/o Sim2Real-Domain-LoRA' row is a model trained entirely without ϕD, which is a different configuration; and it is evaluated only on the CG test set where the domain gap is absent. Fig. 5 demonstrates appearance preservation only, not motion-transfer equivalence. An ablation on a photorealistic benchmark (e.g., NeRSemble) comparing HairWeaver to the same architecture trained without ϕD—with identical conditions and training data—is required to establish that discarding ϕD is harmless and that the two-stage strategy is responsible for the reported gains.
  2. [§4.1 and §9 (Tables 2, 6, 7)] The NeRSemble evaluation changes two variables simultaneously: the conditioning signals (alpha/DensePose instead of UVW/normal) and the training data (additional TikTok training). Consequently, Table 2 cannot isolate the contribution of the Sim2Real-Domain-LoRA mechanism, nor the contribution of the UVW hair condition. The comparison would be more convincing with a controlled variant on NeRSemble using the same condition representation and the same training data across methods, or at least an ablation that isolates each variable.
  3. [§4.2, Table 1] On the self-collected CG test set, HairWeaver is given ground-truth hair UVW conditions rendered from the same simulator that produced the target videos, while the baselines receive no hair-specific condition. This is an apples-to-oranges comparison: the large metric advantages may partly reflect the oracle nature of the hair condition rather than the method's generative superiority. Please either provide baselines with an equivalent hair-conditioning signal (e.g., ControlHair, or a pose-only variant of HairWeaver) or report results on a test set where HairWeaver does not receive simulator-derived UVW maps, so that the state-of-the-art claim is not confounded by privileged conditioning.
minor comments (5)
  1. [Abstract and §3.4] The abstract calls the second module 'Style-Alignment-LoRA,' but the method section consistently refers to 'Sim2Real-Domain-LoRA.' Please unify the terminology.
  2. [§4.2 vs §7, Tables 4–6] The user study description states 50 users in §4.2, but §7 and Table 5 state 30 participants, and Table 6 reports 45 participants. The inconsistency should be corrected and the final participant count confirmed.
  3. [Table 1 caption] Specify whether the hair-area metrics computed with the Matte-Anything mask are evaluated on the generated frames only, on the ground-truth frames, or both, and whether the mask is applied before computing FID/cd-FVD.
  4. [§9] The sentence 'we directly use the alpha channel of hair UVW images as hair condition' is unclear because UVW images are not described as containing an alpha channel. Please clarify the exact channel structure used.
  5. [§2.2 and References] The reference to 'Houdini' appears at the start of the reference list but is not cited in the running text; the footnote-style link in §2.2 is nonstandard. Please cite it properly or remove it.

Circularity Check

0 steps flagged

No significant circularity: the two-stage sim-to-real recipe is an empirical, under-ablated hypothesis, not a derivation that reduces to its inputs.

full rationale

Walking the derivation chain (Eqs. 7-9), I find no step in which a predicted quantity is defined in terms of the same quantity, a fitted parameter is renamed as a prediction, or a load-bearing premise is imported solely from the authors' prior work. The two-stage training is an empirical recipe: Stage 1 learns phi_D on CG I2V (Eq. 7); Stage 2 freezes phi_D and trains phi_P on the full quadruplet (Eq. 8); inference removes phi_D (Eq. 9). The gap between training and inference configuration is real and under-tested, but it is a hypothesis, not a tautology: nothing in Eqs. 7-9 forces epsilon_{theta,phi_P} to inherit the motion mapping learned inside epsilon_{theta,phi_D,phi_P}. The paper's own Ablation Analysis admits the quantitative ablation cannot observe the domain-gap benefit ('Since the self-collected test set has ground-truth video from the CG simulator domain, it is challenging to observe the improvement from Sim2Real-Domain-LoRA in overcoming the domain gap between the CG training data and the photorealistic test data'), and the NeRSemble results use extra TikTok training with different conditions (Section 9), so the sim-to-real contribution is not cleanly isolated. Those are experimental-support limitations, not circularity. The CG test set gives HairWeaver a dense, simulator-derived hair condition (UVW) that is unavailable to baselines, which explains part of the margin, but conditional generation from a user-provided driving signal is the stated task rather than a hidden circular prediction. Self-citations are confined to related work and baseline references; none carries the derivation. Verdict: no significant circularity.

Axiom & Free-Parameter Ledger

3 free parameters · 4 axioms · 3 invented entities

The central claim rests on the CG dataset being a sufficient motion teacher, on the frozen VAE encoder being able to digest UVW/alpha maps as if they were images, and on the motion-LoRA being domain-invariant after the temporary domain-LoRA is discarded. None of these is proven by a formal derivation; they are empirical assumptions validated only indirectly by ablations.

free parameters (3)
  • Learning rate (AdamW) = 2e-4
    Chosen by hand for both training stages; no sweep or sensitivity analysis reported (§4.1).
  • Training steps per stage = 10,000 / 10,000
    Both stages run for 10k steps; no convergence criterion or ablation of step count is given (§4.1).
  • SDEdit noising timestep for photoreal reference generation = 0.35
    Used to balance alignment vs realism when converting CG frames to photoreal references; no ablation of this value is provided (§8).
axioms (4)
  • domain assumption Cosserat-rod physics simulation produces physically plausible and sufficiently diverse hair motion to train a generalizable motion controller.
    The entire synthetic-data pipeline in §3.2 rests on this; no real-data validation of simulation realism is provided.
  • ad hoc to paper A VAE encoder trained on natural RGB video can faithfully encode UVW hair maps and alpha/densepose proxies into tokens the DiT can interpret.
    Equation (5) passes condition maps through the frozen VAE encoder; no analysis addresses the channel-domain mismatch between RGB and UVW.
  • ad hoc to paper Motion control learned while the Sim2Real-Domain-LoRA is frozen remains intact after that LoRA is discarded.
    This is the core two-stage claim: Eq. (8) trains ϕP with ϕD, Eq. (9) removes ϕD. It is never directly ablated.
  • domain assumption Alpha maps plus densepose are sufficient substitutes for UVW plus normal conditions on NeRSemble without retraining the motion pathway.
    Section 9 changes the condition modality at test time; no quantitative study verifies that the motion controller transfers across these very different inputs.
invented entities (3)
  • Motion-Context-LoRA no independent evidence
    purpose: Injects body pose and hair UVW conditions into the frozen DiT via residual pose tokens and concatenated hair tokens.
    Supported only by within-paper ablations; no released weights or external benchmark.
  • Sim2Real-Domain-LoRA no independent evidence
    purpose: Temporarily adapts the backbone to the CG domain during training and is discarded at inference.
    The discard strategy is not directly validated; Fig. 5 shows appearance preservation but not motion-transfer equivalence.
  • UVW strand-map conditioning channel no independent evidence
    purpose: A dense per-pixel hair motion representation from the simulator, used as the hair condition.
    Its transfer to real video is only indirectly evidenced through the alpha-map proxy on NeRSemble.

pith-pipeline@v1.3.0-alltime-deepseek · 15381 in / 12767 out tokens · 125609 ms · 2026-08-03T00:11:40.318117+00:00 · methodology

0 comments
read the original abstract

We present HairWeaver, a diffusion-based pipeline that animates a single human image with realistic and expressive hair dynamics. While existing methods successfully control body pose, they lack specific control over hair, and as a result, fail to capture the intricate hair motions, resulting in stiff and unrealistic animations. HairWeaver overcomes this limitation using two specialized modules: a Motion-Context-LoRA to integrate motion conditions and a Style-Alignment-LoRA to preserve the subject's photoreal appearance across different data domains. These lightweight components are designed to guide a video diffusion backbone while maintaining its core generative capabilities. By training on a specialized dataset of dynamic human motion generated from a CG simulator, HairWeaver affords fine control over hair motion and ultimately learns to produce highly realistic hair that responds naturally to movement. Comprehensive evaluations demonstrate that our approach sets a new state of the art, producing lifelike human hair animations with dynamic details.

discussion (0)

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

Forward citations

Cited by 1 Pith paper

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

  1. MoZoo:Unleashing Video Diffusion power in animal fur and muscle simulation

    cs.GR 2026-04 unverdicted novelty 7.0

    MoZoo generates high-fidelity animal videos with fur and muscle dynamics from coarse meshes by extending video diffusion with role-aware RoPE and asymmetric decoupled attention, trained on a new synthetic-to-real dataset.

Reference graph

Works this paper leans on

23 extracted references · 1 canonical work pages · cited by 1 Pith paper

  1. [1]

    Accessed: 2025-11-12

    Houdini.https://www.sidefx.com/. Accessed: 2025-11-12. Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, et al. Qwen2. 5-vl technical report.arXiv preprint arXiv:2502.13923,

    Pith/arXiv arXiv 2025

  2. [3]

    X-dancer: Expressive music to human dance video generation.arXiv preprint arXiv:2502.17414,

    Zeyuan Chen, Hongyi Xu, Guoxian Song, You Xie, Chenxu Zhang, Xin Chen, Chao Wang, Di Chang, and Linjie Luo. X-dancer: Expressive music to human dance video generation.arXiv preprint arXiv:2502.17414,

    Pith/arXiv arXiv

  3. [4]

    A practical and controllable hair and fur model for production path tracing

    Matt Jen-Yuan Chiang, Benedikt Bitterli, Chuck Tappan, and Brent Burley. A practical and controllable hair and fur model for production path tracing. InACM SIGGRAPH 2015 Talks, pages 1–1

    2015

  4. [9]

    doi: 10.1145/3592455.https://doi.org/10.1145/3592455

    ISSN 0730-0301. doi: 10.1145/3592455.https://doi.org/10.1145/3592455. Tassilo Kugelstadt and Elmar Schömer. Position and orientation based cosserat rods. InSymposium on Computer Animation, volume 11, pages 169–178,

    work page doi:10.1145/3592455.https://doi.org/10.1145/3592455

  5. [10]

    Flux.1 kontext: Flow matching for in-context image generation and editing in latent space, 2025.https://arxiv.org/abs/2506.15742

    14 Black Forest Labs, Stephen Batifol, Andreas Blattmann, Frederic Boesel, Saksham Consul, Cyril Diagne, Tim Dockhorn, Jack English, Zion English, Patrick Esser, Sumith Kulal, Kyle Lacey, Yam Levi, Cheng Li, Dominik Lorenz, Jonas Müller, Dustin Podell, Robin Rombach, Harry Saini, Axel Sauer, and Luke Smith. Flux.1 kontext: Flow matching for in-context ima...

    Pith/arXiv arXiv 2025

  6. [11]

    doi: 10.1109/VR55154.2023.000XX

    IEEE. doi: 10.1109/VR55154.2023.000XX. Yue Li, Gene Wei-Chin Lin, Egor Larionov, Aljaz Bozic, Doug Roble, Ladislav Kavan, Stelian Coros, Bernhard Thomaszewski, Tuur Stuyck, and Hsiao-Yu Chen. Self-supervised learning of latent space dynamics.Proceedings of the ACM on Computer Graphics and Interactive Techniques, 8(4):1–18,

    arXiv 2023

  7. [12]

    Neuralocks: Real-time dynamic neural hair simulation.arXiv preprint arXiv:2507.05191, 2025a

    Gene Wei-Chin Lin, Egor Larionov, Hsiao-yu Chen, Doug Roble, and Tuur Stuyck. Neuralocks: Real-time dynamic neural hair simulation.arXiv preprint arXiv:2507.05191, 2025a. Weikai Lin, Haoxiang Li, and Yuhao Zhu. Controlhair: Physically-based video diffusion for controllable dynamic hair rendering.arXiv preprint arXiv:2509.21541, 2025b. Yuxuan Luo, Zhengkun...

    Pith/arXiv arXiv

  8. [14]

    Towards accurate generative models of video: A new metric & challenges.arXiv preprint arXiv:1812.01717,

    Thomas Unterthiner, Sjoerd Van Steenkiste, Karol Kurach, Raphael Marinier, Marcin Michalski, and Sylvain Gelly. Towards accurate generative models of video: A new metric & challenges.arXiv preprint arXiv:1812.01717,

    Pith/arXiv arXiv

  9. [15]

    Wan: Open and advanced large-scale video generative models.arXiv preprint arXiv:2503.20314,

    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

  10. [16]

    Dreamactor-h1: High-fidelity human-product demonstration video generation via motion-designed diffusion transformers.arXiv preprint arXiv:2506.10568, 2025a

    Lizhen Wang, Zhurong Xia, Tianshu Hu, Pengrui Wang, Pengfei Wei, Zerong Zheng, Ming Zhou, Yuan Zhang, and Mingyuan Gao. Dreamactor-h1: High-fidelity human-product demonstration video generation via motion-designed diffusion transformers.arXiv preprint arXiv:2506.10568, 2025a. Qilin Wang, Zhengkai Jiang, Chengming Xu, Jiangning Zhang, Yabiao Wang, Xinyi Zh...

    Pith/arXiv arXiv

  11. [17]

    Disco: Disentangled control for referring human dance generation in real world.arXiv preprint arXiv:2307.00040, 2023a

    15 Tan Wang, Linjie Li, Kevin Lin, Chung-Ching Lin, Zhengyuan Yang, Hanwang Zhang, Zicheng Liu, and Lijuan Wang. Disco: Disentangled control for referring human dance generation in real world.arXiv preprint arXiv:2307.00040, 2023a. Xiang Wang, Shiwei Zhang, Changxin Gao, Jiayu Wang, Xiaoqiang Zhou, Yingya Zhang, Luxin Yan, and Nong Sang. Unianimate: Tamin...

    Pith/arXiv arXiv

  12. [18]

    Ip-adapter: Text compatible image prompt adapter for text-to-image diffusion models.arXiv preprint arXiv:2308.06721,

    Hu Ye, Jun Zhang, Sibo Liu, Xiao Han, and Wei Yang. Ip-adapter: Text compatible image prompt adapter for text-to-image diffusion models.arXiv preprint arXiv:2308.06721,

    Pith/arXiv arXiv

  13. [19]

    X-actor: Emotional and expressive long-range portrait acting from audio.arXiv preprint arXiv:2508.02944, 2025a

    Chenxu Zhang, Zenan Li, Hongyi Xu, You Xie, Xiaochen Zhao, Tianpei Gu, Guoxian Song, Xin Chen, Chao Liang, Jianwen Jiang, et al. X-actor: Emotional and expressive long-range portrait acting from audio.arXiv preprint arXiv:2508.02944, 2025a. Joy Xiaoji Zhang, Jingsen Zhu, Hanyu Chen, and Steve Marschner. Hairformer: Transformer-based dynamic neural hair si...

    Pith/arXiv arXiv

  14. [20]

    Mimicmotion: High-quality human motion video generation with confidence-aware pose guidance.arXiv preprint arXiv:2406.19680,

    Yuang Zhang, Jiaxi Gu, Li-Wen Wang, Han Wang, Junqi Cheng, Yuefeng Zhu, and Fangyuan Zou. Mimicmotion: High-quality human motion video generation with confidence-aware pose guidance.arXiv preprint arXiv:2406.19680,

    Pith/arXiv arXiv

  15. [21]

    X-nemo: Expressive neural motion reenactment via disentangled latent attention.arXiv preprint arXiv:2507.23143,

    Xiaochen Zhao, Hongyi Xu, Guoxian Song, You Xie, Chenxu Zhang, Xiu Li, Linjie Luo, Jinli Suo, and Yebin Liu. X-nemo: Expressive neural motion reenactment via disentangled latent attention.arXiv preprint arXiv:2507.23143,

    Pith/arXiv arXiv

  16. [22]

    Transfusion: Predict the next token and diffuse images with one multi-modal model.arXiv preprint arXiv:2408.11039,

    Chunting Zhou, Lili Yu, Arun Babu, Kushal Tirumala, Michihiro Yasunaga, Leonid Shamis, Jacob Kahn, Xuezhe Ma, Luke Zettlemoyer, and Omer Levy. Transfusion: Predict the next token and diffuse images with one multi-modal model.arXiv preprint arXiv:2408.11039,

    Pith/arXiv arXiv

  17. [23]

    Champ: Controllable and consistent human image animation with 3d parametric guidance.arXiv preprint arXiv:2403.14781,

    Shenhao Zhu, Junming Leo Chen, Zuozhuo Dai, Yinghui Xu, Xun Cao, Yao Yao, Hao Zhu, and Siyu Zhu. Champ: Controllable and consistent human image animation with 3d parametric guidance.arXiv preprint arXiv:2403.14781,

    Pith/arXiv arXiv

  18. [2018]

    Animatediff: Animate your personalized text-to-image diffusion models without specific tuning.arXiv preprint arXiv:2307.04725,

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

    Pith/arXiv arXiv

  19. [2021]

    3d gaussian splatting for real-time radiance field rendering

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimk"uhler, and George Drettakis. 3d gaussian splatting for real-time radiance field rendering. InACM SIGGRAPH 2023 Transactions,

    2023

  20. [2022]

    X- unimotion: Animating human images with expressive, unified and identity-agnostic motion latents.arXiv preprint arXiv:2508.09383,

    Guoxian Song, Hongyi Xu, Xiaochen Zhao, You Xie, Tianpei Gu, Zenan Li, Chenxu Zhang, and Linjie Luo. X- unimotion: Animating human images with expressive, unified and identity-agnostic motion latents.arXiv preprint arXiv:2508.09383,

    Pith/arXiv arXiv

  21. [2023]

    Ltx-video: Realtime video latent diffusion.arXiv preprint arXiv:2501.00103,

    Yoav HaCohen, Nisan Chiprut, Benny Brazowski, Daniel Shalem, Dudu Moshe, Eitan Richardson, Eran Levin, Guy Shiran, Nir Zabari, Ori Gordon, Poriya Panet, Sapir Weissbuch, Victor Kulikov, Yaki Bitterman, Zeev Melumian, and Ofir Bibi. Ltx-video: Realtime video latent diffusion.arXiv preprint arXiv:2501.00103,

    Pith/arXiv arXiv

  22. [2024]

    Songwei Ge, Aniruddha Mahapatra, Gaurav Parmar, Jun-Yan Zhu, and Jia-Bin Huang

    Accessed: 2025-07-17. Songwei Ge, Aniruddha Mahapatra, Gaurav Parmar, Jun-Yan Zhu, and Jia-Bin Huang. On the content bias in fréchet video distance. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR),

    2025

  23. [2025]

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

    Andreas Blattmann, Tim Dockhorn, Sumith Kulal, Daniel Mendelevitch, Maciej Kilian, Dominik Lorenz, Yam Levi, Zion English, Vikram Voleti, Adam Letts, et al. Stable video diffusion: Scaling latent video diffusion models to large datasets.arXiv preprint arXiv:2311.15127, 2023a. Andreas Blattmann, Robin Rombach, Huan Ling, Tim Dockhorn, Seung Wook Kim, Sanja...

    Pith/arXiv arXiv