arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2502.02492cs.CV
收藏

VideoJAM:用于增强视频模型运动生成的联合外观-运动表示

VideoJAM: Joint Appearance-Motion Representations for Enhanced Motion Generation in Video Models

GenAI, Meta(Meta GenAI);Tel Aviv University(特拉维夫大学)

Hila Chefer, Uriel Singer, Amit Zohar, Yuval Kirstain, Adam Polyak, Yaniv Taigman, Lior Wolf, Shelly Sheynin

AI总结:

VideoJAM通过联合外观-运动表示和Inner-Guidance机制,在训练和推理中增强视频生成的运动连贯性,无需修改数据或扩大模型,达到最先进性能并提升视觉质量。

摘要:

尽管近期取得了巨大进展,生成式视频模型仍然难以捕捉真实世界的运动、动态和物理规律。我们表明,这一局限性源于传统的像素重建目标,该目标使模型偏向于外观保真度,而牺牲了运动连贯性。为解决此问题,我们提出了VideoJAM,一种新颖的框架,通过鼓励模型学习联合的外观-运动表示,为视频生成器注入有效的运动先验。VideoJAM由两个互补的单元组成。在训练期间,我们扩展了目标函数,使其从单一学习表示中同时预测生成的像素及其对应的运动。在推理期间,我们引入了Inner-Guidance机制,该机制利用模型自身不断演化的运动预测作为动态引导信号,将生成过程导向连贯的运动。值得注意的是,我们的框架可以以最小的改动应用于任何视频模型,无需修改训练数据或扩大模型规模。VideoJAM在运动连贯性方面达到了最先进的性能,超越了极具竞争力的专有模型,同时增强了生成内容的感知视觉质量。这些发现强调,外观和运动可以是互补的,当有效整合时,能同时提升视频生成的视觉质量和连贯性。项目网站:https://hila-chefer.github.io/videojam-paper.github.io/

英文摘要:

Despite tremendous recent progress, generative video models still struggle to capture real-world motion, dynamics, and physics. We show that this limitation arises from the conventional pixel reconstruction objective, which biases models toward appearance fidelity at the expense of motion coherence. To address this, we introduce VideoJAM, a novel framework that instills an effective motion prior to video generators, by encouraging the model to learn a joint appearance-motion representation. VideoJAM is composed of two complementary units. During training, we extend the objective to predict both the generated pixels and their corresponding motion from a single learned representation. During inference, we introduce Inner-Guidance, a mechanism that steers the generation toward coherent motion by leveraging the model's own evolving motion prediction as a dynamic guidance signal. Notably, our framework can be applied to any video model with minimal adaptations, requiring no modifications to the training data or scaling of the model. VideoJAM achieves state-of-the-art performance in motion coherence, surpassing highly competitive proprietary models while also enhancing the perceived visual quality of the generations. These findings emphasize that appearance and motion can be complementary and, when effectively integrated, enhance both the visual quality and the coherence of video generation. Project website: https://hila-chefer.github.io/videojam-paper.github.io/

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑