arXiv 2507.112612025-07-27cs.CV

ViewSRD：通过结构化多视图分解实现3D视觉定位

ViewSRD: 3D Visual Grounding via Structured Multi-View Decomposition

South China University of Technology（华南理工大学）；Guangdong Engineering Center for Large Model and GenAI Technology（广东省大模型与生成式人工智能技术工程中心）；State Key Laboratory of Subtropical Building and Urban Science（亚热带建筑科学国家重点实验室）；Ministry of Education Key Laboratory of Big Data and Intelligent Robot（教育部大数据与智能机器人重点实验室）；Guangdong Provincial Key Lab of Computational Intelligence and Cyberspace Information（广东省计算智能与网络信息重点实验室）；Singapore Management University（新加坡国立大学）

Ronggang Huang, Haoxin Yang, Yan Cai, Xuemiao Xu, Huaidong Zhang, Shengfeng He

备注：: Accepted by ICCV 2025

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出ViewSRD框架，通过结构化多视图分解、跨视图文本-场景交互与推理，解决复杂多锚点查询中的目标解耦及视角不一致问题。

摘要：

3D视觉定位旨在根据文本描述在3D空间中识别并定位物体。然而，现有方法在复杂的多锚点查询中难以将目标与锚点解耦，也难以解决由视角变化引起的空间描述不一致问题。为应对这些挑战，我们提出ViewSRD，该框架将3D视觉定位表述为结构化多视图分解过程。首先，Simple Relation Decoupling（SRD，简单关系解耦）模块将复杂的多锚点查询重构为一组有针对性的单锚点陈述，生成一组结构化的视角感知描述，以明确位置关系。这些分解后的表示作为Multi-view Textual-Scene Interaction（Multi-TSI，多视图文本-场景交互）模块的基础，该模块使用共享的Cross-modal Consistent View Tokens（CCVT，跨模态一致视图令牌）跨多个视角整合文本与场景特征，以保留空间相关性。最后，Textual-Scene Reasoning（文本-场景推理）模块将多视图预测合成为统一且鲁棒的3D视觉定位结果。在3D视觉定位数据集上的实验表明，ViewSRD显著优于最先进方法，尤其是在需要精确空间区分的复杂查询中。代码位于https://github.com/visualjason/ViewSRD。

英文摘要：

3D visual grounding aims to identify and localize objects in a 3D space based on textual descriptions. However, existing methods struggle with disentangling targets from anchors in complex multi-anchor queries and resolving inconsistencies in spatial descriptions caused by perspective variations. To tackle these challenges, we propose ViewSRD, a framework that formulates 3D visual grounding as a structured multi-view decomposition process. First, the Simple Relation Decoupling (SRD) module restructures complex multi-anchor queries into a set of targeted single-anchor statements, generating a structured set of perspective-aware descriptions that clarify positional relationships. These decomposed representations serve as the foundation for the Multi-view Textual-Scene Interaction (Multi-TSI) module, which integrates textual and scene features across multiple viewpoints using shared, Cross-modal Consistent View Tokens (CCVTs) to preserve spatial correlations. Finally, a Textual-Scene Reasoning module synthesizes multi-view predictions into a unified and robust 3D visual grounding. Experiments on 3D visual grounding datasets show that ViewSRD significantly outperforms state-of-the-art methods, particularly in complex queries requiring precise spatial differentiation. Code is available at https://github.com/visualjason/ViewSRD.

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。