复制-变换-粘贴：视觉语言与几何约束引导的零样本物体-物体对齐
Copy-Transform-Paste: Zero-Shot Object-Object Alignment Guided by Vision-Language and Geometric Constraints
提出一种无需训练的测试时优化框架，利用 CLIP 梯度与可微渲染直接优化两个网格的相对位姿，并引入软 ICP 与穿透损失等几何约束，可复用于语言引导的三维场景组装与物体摆放任务。
原文依据 we directly optimize the relative pose at test time---updating translation, rotation, and isotropic scale with CLIP-driven gradients via a differentiable renderer---without training a new model.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论我们研究在简短文本提示描述两个给定网格空间关系的情况下，实现零样本三维对齐——这是内容创作和场景组装的一项基本能力。早期方法主要依赖几何对齐流程，而近期工作利用预训练的二维扩散模型来建模语言条件化的物体-物体空间关系。相比之下，我们在测试时直接优化相对位姿——通过可微渲染器，利用 CLIP 驱动的梯度更新平移、旋转和各向同性缩放——而无需训练新模型。我们的框架通过几何感知目标增强语言监督：一种软迭代最近点（ICP）项的变体，用于鼓励受控的表面附着，以及一个穿透损失，用于抑制相互穿插。分阶段调度随时间增强接触约束，相机控制将视图集中在交互区域，随机重启提升鲁棒性。为便于评估，我们整理了一个包含 50 个网格对-提示词案例的基准，涵盖多样类别和关系，并与基线方法进行比较。在该基准上，我们的方法产生语义忠实且物理合理的对齐，在提高 CLIP 相似度的同时减少相交体积。
英文原文摘要 1283 字符
We study zero-shot 3D alignment of two given meshes from a short text prompt describing their spatial relation---an essential capability for content creation and scene assembly. Earlier approaches primarily rely on geometric alignment procedures, while recent work leverages pretrained 2D diffusion models to model language-conditioned object-object spatial relationships. In contrast, we directly optimize the relative pose at test time---updating translation, rotation, and isotropic scale with CLIP-driven gradients via a differentiable renderer---without training a new model.Our framework augments language supervision with geometry-aware objectives: a variant of soft-Iterative Closest Point (ICP) term to encourage controlled surface attachment and a penetration loss to discourage interpenetration. A phased schedule strengthens contact constraints over time, camera control concentrates views on the interaction region, and randomized restarts improve robustness.To enable evaluation, we curate a benchmark of 50 \ mesh pair, prompt\ cases spanning diverse categories and relations, and compare against baselines. Across the benchmark, our method yields semantically faithful and physically plausible alignments, improving CLIP similarity while reducing intersection volume.