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 2409.17686cs.CV
收藏

MoGenTS:基于时空联合建模的运动生成

MoGenTS: Motion Generation based on Spatial-Temporal Joint Modeling

Alibaba Group(阿里巴巴集团);The University of Texas at Austin(德克萨斯大学奥斯汀分校)

Weihao Yuan, Weichao Shen, Yisheng He, Yuan Dong, Xiaodong Gu, Zilong Dong, Liefeng Bo, Qixing Huang

AI总结:

提出MoGenTS框架,通过将每个独立关节量化为向量以保留时空结构,并利用2D关节VQVAE、时空2D掩码与注意力机制建模时空信号,在多个数据集上显著提升运动生成性能。

摘要:

基于离散量化的运动生成相比连续回归具有许多优势,但代价是不可避免地存在近似误差。以往的方法通常将整个身体姿态量化为一个码本,这不仅面临将所有关节编码进单一向量的困难,还丢失了不同关节之间的空间关系。不同的是,在本工作中,我们将每个独立的关节量化为一个向量,这 i) 简化了量化过程,因为与单一关节相关的复杂性明显低于整个姿态;ii) 保持了时空结构,从而保留了关节间的空间关系和时间运动模式;iii) 产生了一个2D token map,使得能够应用2D图像中广泛使用的各种2D操作。基于2D运动量化,我们构建了一个时空建模框架,其中提出了2D关节VQVAE、时空2D掩码技术以及时空2D注意力,以利用2D tokens之间的时空信号。大量实验表明,我们的方法在不同数据集上均显著优于以往方法,在HumanML3D上FID降低了26.6%,在KIT-ML上降低了29.9%。项目页面:https://aigc3d.github.io/mogents。

英文摘要:

Motion generation from discrete quantization offers many advantages over continuous regression, but at the cost of inevitable approximation errors. Previous methods usually quantize the entire body pose into one code, which not only faces the difficulty in encoding all joints within one vector but also loses the spatial relationship between different joints. Differently, in this work we quantize each individual joint into one vector, which i) simplifies the quantization process as the complexity associated with a single joint is markedly lower than that of the entire pose; ii) maintains a spatial-temporal structure that preserves both the spatial relationships among joints and the temporal movement patterns; iii) yields a 2D token map, which enables the application of various 2D operations widely used in 2D images. Grounded in the 2D motion quantization, we build a spatial-temporal modeling framework, where 2D joint VQVAE, temporal-spatial 2D masking technique, and spatial-temporal 2D attention are proposed to take advantage of spatial-temporal signals among the 2D tokens. Extensive experiments demonstrate that our method significantly outperforms previous methods across different datasets, with a 26.6% decrease of FID on HumanML3D and a 29.9% decrease on KIT-ML. Project page: https://aigc3d.github.io/mogents.

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。

↑