ORD：面向泛化 3D 视觉定位的对象-关系解耦
ORD: Object-Relation Decoupling for Generalized 3D Visual Grounding
将目标-锚框关系显式解耦为几何与语义原语，并设计谓词解耦的跨模态对齐策略，可迁移到其他需要抑制实体名称语义捷径的视觉-语言任务（如指代分割、视觉定位）。
原文依据 we propose an object-relation decoupling framework that treats target-anchor relations as first-class geometric and semantic primitives and models them explicitly.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论3D 视觉定位任务旨在根据自然语言描述准确识别并定位 3D 空间中的目标对象，其中有效利用目标与锚框之间的相对关系至关重要。然而，现有方法中相对关系往往与实体语义紧密纠缠。这种紧耦合促使模型依赖实体名称带来的语义捷径，使其在多视图和复杂多对象场景下难以保持良好的泛化能力。为解决这一问题，我们提出一种对象-关系解耦框架，将目标-锚框关系视为一等的几何与语义原语并对其进行显式建模。首先，我们构建场景级相对几何表示，编码目标与锚框之间的方向和距离，并引入场景级超对象词元作为尺度和视角的统一先验。其次，我们开发谓词解耦的跨模态对齐策略，仅保留携带空间关系语义的谓词，同时掩蔽所有其他词元，从而抑制实体名称的语义泄漏。最后，我们设计锚框引导的回归模块，预测辅助锚框并采样其特征以引导模型从文本中学习实体语义，显式注入目标-锚框先验，有效解决复杂多对象场景中的歧义。在多个 3D 视觉定位基准上的大量实验表明，我们的方法持续优于最优方法，并在具有挑战性的多视图和关系密集设置下展现出强鲁棒性和泛化能力。
英文原文摘要 1701 字符
3D visual grounding task aims to accurately identify and grounding target objects in 3D space based on natural language descriptions,where the effective exploitation of relative relations between the target and anchor is crucial.However, in existing methods, relative relations are often tightly entangled with entity semantics. This tight coupling encourages models to rely on semantic shortcuts from entity names, making it difficult to maintain good generalization under multi-view and complex multi-object scenarios.To address this, we propose an object-relation decoupling framework that treats target-anchor relations as first-class geometric and semantic primitives and models them explicitly.First, we construct a scene-level relative geometric representation that encodes the direction and distance between the target and anchor, and introduce a scene-level hyper-object token as a unified prior for scale and viewpoint.Second, we develop a predicate-decoupled cross-modal alignment strategy that preserves only predicates carrying spatial relational semantics while masking out all other tokens, thereby suppressing semantic leakage from entity names.Finally, we design an anchor-guided regression module that predicts auxiliary anchors and samples their features to guide the model in learning entity semantics from text, explicitly injecting target-anchor priors and effectively resolving ambiguities in complex multi-object scenes.Extensive experiments on multiple 3D visual grounding benchmarks demonstrate that our method consistently outperforms state-of-the-art approaches and exhibits strong robustness and generalization under challenging multi-view and relation-intensive settings.