arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2506.18655cs.CV
收藏

RDPO:用于物理一致性视频生成的真实数据偏好优化

RDPO: Real Data Preference Optimization for Physics Consistency Video Generation

Fudan University(复旦大学);Shopee Inc(Shopee公司)

Wenxu Qian, Chaoyue Wang, Hou Peng, Zhiyu Tan, Hao Li, Anxiang Zeng

AI总结:

RDPO提出无需标注的框架,通过从真实视频反向采样构建偏好对并多阶段迭代训练,提升视频生成的物理一致性与动作连贯性。

摘要:

视频生成技术在视觉质量方面取得了显著进步,但忠实再现真实世界的物理规律仍然难以实现。基于偏好的模型后训练可能改善物理一致性,但需要昂贵的人工标注数据集或尚不可行的奖励模型。为解决这些挑战,我们提出了真实数据偏好优化(RDPO),一种无需标注的框架,直接从真实世界视频中提炼物理先验。具体而言,所提出的RDPO利用预训练生成器对真实视频序列进行反向采样,自动构建在物理正确性方面具有统计可区分性的偏好对。随后,多阶段迭代训练计划引导生成器日益更好地遵循物理定律。得益于从真实视频中挖掘的动态信息,我们提出的RDPO显著提升了生成视频的动作连贯性和物理真实性。在多个基准上的评估和人工评估表明,RDPO在多个维度上实现了改进。本文的源代码和演示可在以下网址获取:https://wwenxu.github.io/RDPO/

英文摘要:

Video generation techniques have achieved remarkable advancements in visual quality, yet faithfully reproducing real-world physics remains elusive. Preference-based model post-training may improve physical consistency, but requires costly human-annotated datasets or reward models that are not yet feasible. To address these challenges, we present Real Data Preference Optimisation (RDPO), an annotation-free framework that distills physical priors directly from real-world videos. Specifically, the proposed RDPO reverse-samples real video sequences with a pre-trained generator to automatically build preference pairs that are statistically distinguishable in terms of physical correctness. A multi-stage iterative training schedule then guides the generator to obey physical laws increasingly well. Benefiting from the dynamic information explored from real videos, our proposed RDPO significantly improves the action coherence and physical realism of the generated videos. Evaluations on multiple benchmarks and human evaluations have demonstrated that RDPO achieves improvements across multiple dimensions. The source code and demonstration of this paper are available at: https://wwenxu.github.io/RDPO/

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。

↑