Skip to content
Preprint

QVR3D: Query-Guided Visual Reasoning for Training-Free 3D Grounding

Oct 2026 · 0 citations · 54 references
Computer Science

TL;DR

QVR3D is introduced, a query-guided visual reasoning framework that operates on RGB-D observations using off-the-shelf foundation models that integrates instance confidence, attribute consistency, and relationship consistency and supports both single- and multi-object grounding without task-specific adaptation.

Abstract

3D visual grounding localizes objects in 3D scenes from natural language descriptions, requiring reasoning over semantics, spatial relationships, and multi-view observations. Training-free approaches avoid costly annotations but often rely on explicit scene transformation or scene-level frame selection, which may waste limited visual capacity on query-irrelevant regions. We observe that effective grounding should prioritize observations that provide evidence for the query rather than uniformly represent the entire scene. Based on this insight, we introduce QVR3D, a query-guided visual reasoning framework that operates on RGB-D observations using off-the-shelf foundation models. QVR3D combines Query-Guided Maximum Coverage (QGMC) for selecting complementary query-relevant views with Grounding Probability Estimation (GPE) that integrates instance confidence, attribute consistency, and relationship consistency. The framework supports both single- and multi-object grounding without task-specific adaptation. Experiments on ScanRefer and Nr3D achieve state-of-the-art zero-shot performance, while QGMC also improves frame efficiency across multiple 3D vision-language tasks. Our code is available at https://github.com/jwliao1209/QVR3D.

View source

Similar papers

Preprint Oct 2026

Grounding Across Views: Cross-View Referring Reasoning for Visual Grounding

Cross-view object geo-localization (CVOGL) aims to localize a target in a satellite image from an observation captured under another viewpoint. Existing methods typically specify the query with a click point or bounding box, marking the target region but underusing the semantic attributes and relations conveyed by natu...

Hao-Xiang Gan, Liu-Qiang Wang, Shuang Li et al. · 0 citations
Nov 2026

SAOG: Spatial-Attribute-Aware 3D Object Grounding via Visual Question Answering-Enhanced Reasoning

3D object grounding aims to identify the target object referred to by natural language instructions in the presence of multiple distractors. The task is inherently challenging, as it demands joint consideration of spatial relations and intrinsic object attributes to disambiguate the referred target from distractors. Co...

Tong-Xing Jin, Jie Xu, Sheng-Hai Yuan et al. · 1 citation
Preprint Nov 2026

BEV3SR: Evidence-Aware 3D Spatial Reasoning from Video

3D spatial reasoning from video remains challenging for VLMs because videos provide sequential, view-local observations rather than persistent metric scene records. Many questions must be answered in a query-specified perspective rather than the observed camera viewpoint, and prompting with a fixed set of RGB views may...

Fang-Bo Zhu, Rui-Han Yu · 0 citations
Preprint Oct 2026

RelateGS: Reliable Semantic and Relational Evidence for 3D Scene Understanding with Gaussian Splatting

RelateGS is presented, a structured language-aware Gaussian representation that selectively refines semantic and relational evidence for 3D visual grounding and suppresses unreliable observations during semantic feature uplifting and consolidates multi-view object observations into a persistent 3D graph.

Seongjun Choi, Sangyoun Lee · 0 citations
Open access Sep 2026

Spatially Constrained 3D Scene Graphs for Scene-Grounded Subtask Generation

A feed-forward, spatially constrained framework that grounds task-relevant objects and generates position-aware subtasks without an iterative task–scene refinement loop is proposed and confidence-guided verification improves the unweighted scene-average Relaxed Accuracy F1.

Junsang Ryu, Si-Woo Lee, Inseon Choi et al. · 0 citations

Toward Comprehensive 3D Grounding: Orientation Grounding through Vision-Language Models

Ori, a referring grounding task that predicts an object's 6D orientation and axial symmetry from a language or box query in single-view or multi-view scenes, substantially outperforms orientation-aware VLM baselines and surpasses object-level orientation foundation models on scene-level referring benchmarks, showing th...

Tuo Liang, Di-Sheng Liu, Neng-Bo Wang et al. · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.