arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2212.02500cs.CVcs.AIcs.GRcs.LG
收藏

PhysDiff:物理引导的人体运动扩散模型

PhysDiff: Physics-Guided Human Motion Diffusion Model

NVIDIA(英伟达)

Ye Yuan, Jiaming Song, Umar Iqbal, Arash Vahdat, Jan Kautz

AI总结:

针对现有运动扩散模型忽略物理定律、生成运动存在物理伪影的问题,提出PhysDiff模型,通过物理运动投影模块迭代引导扩散过程,实现最先进的运动质量,大幅提升物理可信度。

摘要:

去噪扩散模型在生成多样且逼真的人体运动方面具有广阔前景。然而,现有运动扩散模型在扩散过程中很大程度上忽略了物理定律,往往生成物理上不可信的运动,伴随明显的伪影,如漂浮、脚部滑动和地面穿透。这严重影响了生成运动的质量,并限制了其实际应用。为解决这一问题,我们提出了一种新颖的物理引导运动扩散模型(PhysDiff),将物理约束融入扩散过程。具体而言,我们提出了一个基于物理的运动投影模块,该模块在物理模拟器中使用运动模仿,将扩散步骤的去噪运动投影为物理上可信的运动。投影后的运动将进一步用于下一个扩散步骤,以引导去噪扩散过程。直观地说,我们模型中物理的使用将运动迭代地拉向物理可信空间,这是简单后处理无法实现的。在大规模人体运动数据集上的实验表明,我们的方法实现了最先进的运动质量,并大幅提升了物理可信度(所有数据集提升幅度均>78%)。

英文摘要:

Denoising diffusion models hold great promise for generating diverse and realistic human motions. However, existing motion diffusion models largely disregard the laws of physics in the diffusion process and often generate physically-implausible motions with pronounced artifacts such as floating, foot sliding, and ground penetration. This seriously impacts the quality of generated motions and limits their real-world application. To address this issue, we present a novel physics-guided motion diffusion model (PhysDiff), which incorporates physical constraints into the diffusion process. Specifically, we propose a physics-based motion projection module that uses motion imitation in a physics simulator to project the denoised motion of a diffusion step to a physically-plausible motion. The projected motion is further used in the next diffusion step to guide the denoising diffusion process. Intuitively, the use of physics in our model iteratively pulls the motion toward a physically-plausible space, which cannot be achieved by simple post-processing. Experiments on large-scale human motion datasets show that our approach achieves state-of-the-art motion quality and improves physical plausibility drastically (>78% for all datasets).

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。

↑