arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2403.15249cs.CVcs.AIcs.LG
收藏

基于扩散模型的视频运动迁移的频谱运动对齐

Spectral Motion Alignment for Video Motion Transfer using Diffusion Models

Korea Advanced Institute of Science and Technology (KAIST)(韩国科学技术院)

Geon Yeong Park, Hyeonho Jeong, Sang Wan Lee, Jong Chul Ye

AI总结:

针对视频运动迁移中连续帧残差缺乏全局运动上下文且易受逐帧扭曲影响的问题,提出频谱运动对齐(SMA)框架,利用傅里叶和小波变换细化对齐运动向量并引入频域正则化,有效提升了运动迁移质量与计算效率。

摘要:

扩散模型的演进极大地影响了视频生成与理解。特别是,文本到视频扩散模型(VDMs)显著促进了输入视频在目标外观、运动等方面的定制。尽管取得了这些进展,从视频帧中准确提取运动信息仍存在挑战。虽然现有工作利用连续帧残差作为目标运动向量,但它们本质上缺乏全局运动上下文,且容易受到逐帧扭曲的影响。为了解决这一问题,我们提出了频谱运动对齐(SMA),这是一种利用傅里叶和小波变换来细化和对齐运动向量的新框架。SMA通过引入频域正则化来学习运动模式,促进整帧全局运动动态的学习,并缓解空间伪影。大量实验表明,SMA在改善运动迁移方面具有有效性,同时保持了计算效率以及与各种视频定制框架的兼容性。

英文摘要:

The evolution of diffusion models has greatly impacted video generation and understanding. Particularly, text-to-video diffusion models (VDMs) have significantly facilitated the customization of input video with target appearance, motion, etc. Despite these advances, challenges persist in accurately distilling motion information from video frames. While existing works leverage the consecutive frame residual as the target motion vector, they inherently lack global motion context and are vulnerable to frame-wise distortions. To address this, we present Spectral Motion Alignment (SMA), a novel framework that refines and aligns motion vectors using Fourier and wavelet transforms. SMA learns motion patterns by incorporating frequency-domain regularization, facilitating the learning of whole-frame global motion dynamics, and mitigating spatial artifacts. Extensive experiments demonstrate SMA's efficacy in improving motion transfer while maintaining computational efficiency and compatibility across various video customization frameworks.

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。

↑