arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2501.07563cs.CV
收藏

无需训练的运动引导视频生成:利用运动一致性损失增强时间一致性

Training-Free Motion-Guided Video Generation with Enhanced Temporal Consistency Using Motion Consistency Loss

The University of Adelaide(阿德莱德大学);The University of New South Wales(新南威尔士大学)

Xinyu Zhang, Zicheng Duan, Dong Gong, Lingqiao Liu

AI总结:

针对免训练运动引导视频生成的时间一致性差、运动跟随不准问题,提出结合初始噪声方法与运动一致性损失的方案,通过捕获参考视频帧间特征相关模式、用损失梯度引导生成,在免训练前提下提升了时间一致性,树立了高效连贯视频生成的新基准。

摘要:

本文致力于解决在运动引导下生成具备时间一致性视频的挑战。现有许多方法依赖额外的控制模块或推理阶段微调,而近期研究表明,无需修改模型架构或进行额外训练,也能实现有效的运动引导。这类方法与各类视频生成基础模型的兼容性颇具潜力,但现有免训练方法往往难以在帧间维持稳定的时间连贯性,也无法精准遵循引导运动。\n本研究提出了一种简洁却有效的方案,将基于初始噪声的方法与全新的运动一致性损失相结合,其中运动一致性损失是我们的核心创新。具体而言,我们捕获视频扩散模型中间特征的帧间特征相关模式,用以表征参考视频的运动模式;随后设计运动一致性损失,使生成视频维持相似的特征相关模式,并利用该损失在潜空间中的梯度引导生成过程,实现精准的运动控制。该方法在各类运动控制任务中提升了时间一致性,同时保留了免训练设置的优势。大量实验表明,我们的方法为高效、时间连贯的视频生成树立了新的基准。

英文摘要:

In this paper, we address the challenge of generating temporally consistent videos with motion guidance. While many existing methods depend on additional control modules or inference-time fine-tuning, recent studies suggest that effective motion guidance is achievable without altering the model architecture or requiring extra training. Such approaches offer promising compatibility with various video generation foundation models. However, existing training-free methods often struggle to maintain consistent temporal coherence across frames or to follow guided motion accurately. In this work, we propose a simple yet effective solution that combines an initial-noise-based approach with a novel motion consistency loss, the latter being our key innovation. Specifically, we capture the inter-frame feature correlation patterns of intermediate features from a video diffusion model to represent the motion pattern of the reference video. We then design a motion consistency loss to maintain similar feature correlation patterns in the generated video, using the gradient of this loss in the latent space to guide the generation process for precise motion control. This approach improves temporal consistency across various motion control tasks while preserving the benefits of a training-free setup. Extensive experiments show that our method sets a new standard for efficient, temporally coherent video generation.

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。

↑