[论文] LEXIS: LatEnt ProXimal Interaction Signatures for 3D HOI from an Image

## 论文概要 **研究领域**: CV **作者**: Dimitrije Antić, Alvaro Budria, George Paschalidis **发布时间**: 2026-04-22 **arXiv**: [2604.20800](https://arxiv.org/abs/2604.20800) ## 中文摘要 从RGB图像重建3D人-物交互对感知系统至关重要。然而，这仍然具有挑战性，因为它需要捕捉身体与物体之间微妙的物理耦合。当前方法依赖稀疏的二值接触线索，但这些方法无法建模自然交互中连续邻近和密集空间关系。我们通过InterFields来解决这一局限性，这是一种在整个身体和物体表面编码密集、连续邻近性的表示。然而，从单张图像推断这些场在本质上是病态的。为解决这一问题，我们的直觉是交互模式由动作和物体几何特征性地结构化。我们在LEXIS中捕捉这一结构，这是一种通过VQ-VAE学习的新颖离散流形交互签名。然后我们开发LEXIS-Flow，一种利用LEXIS签名来估计人体和物体网格及其InterFields的扩散框架。值得注意的是，这些InterFields有助于引导细化，确保物理可信、感知邻近性的重建，无需后优化。在Open3DHOI和BEHAVE上的评估表明，LEXIS-Flow在重建、接触和邻近质量方面显著优于现有SotA基线。我们的方法不仅提高了泛化能力，而且产生了被认为更真实的重建，使我们更接近整体3D场景理解。代码和模型将在https://anticdimi.github.io/lexis公开发布。 ## 原文摘要 Reconstructing 3D Human-Object Interaction from an RGB image is essential for perceptive systems. Yet, this remains challenging as it requires capturing the subtle physical coupling between the body and objects. While current methods rely on sparse, binary contact cues, these fail to model the continuous proximity and dense spatial relationships that characterize natural interactions. We address this limitation via InterFields, a representation that encodes dense, continuous proximity across the entire body and object surfaces. However, inferring these fields from single images is inherently ill-posed. To tackle this, our intuition is that interaction patterns are characteristically structured by the action and object geometry. We capture this structure in LEXIS, a novel discrete manifold ... --- *自动采集于 2026-04-24* #论文 #arXiv #CV #小凯