3D Visual Grounding

3D Gaussian Splatting (3DGS) has emerged as a powerful technique for novel view synthesis, subsequently extending into numerous spatial AI applications. However, most existing 3DGS methods operate in isolation, focusing on specific domains. In this paper, we introduce X-GS, an extensible framework that integrates previously isolated 3DGS methods into the perception module of a VLM for spatial tasks, with two major components: the $\textit{Perceiver}$ and the $\textit{Thinker}$. The $\textit{Perceiver}$ performs online 3DGS-based SLAM with semantic distillation and outputs semantic Gaussians from unposed video streams. It leverages recent vision foundation models for stronger geometric priors, and we introduce three novel optimizations to improve semantic distillation efficiency. The $\textit{Thinker}$ interfaces diverse VLMs with these semantic Gaussians, unlocking spatial multimodal capabilities such as 3D visual grounding and scene captioning. Experimental results on diverse benchmarks demonstrate the efficiency and newly unlocked multimodal capabilities of the X-GS framework.

VLA models encode visual observations as 2D patch tokens with no intrinsic geometric structure. We introduce GST-VLA with two contributions. First, the Gaussian Spatial Tokenizer (GST) converts frozen dense depth and frozen semantic patch features into $N_g{=}128$ anisotropic 3D Gaussian primitives, each parameterized by a metric residual mean $μ\in \mathbb{R}^3$, log-scale covariance $\log σ\in \mathbb{R}^3$, and learned opacity $α\in (0,1)$. The covariance eigenstructure encodes local surface orientation, and opacity provides per-primitive geometric confidence, both inaccessible from scalar depth. Spatial attention pooling with learned queries concentrates the fixed token budget on geometrically salient regions rather than distributing uniformly. Second, 3D Depth-Aware Chain-of-Thought (DA-CoT) reasoning supervises four structured intermediate spatial thoughts, covering 3D object grounding, grasp affordance contact geometry, pairwise metric distances, and coarse SE(3) waypoints, as explicit generation targets in the training loss. A cross-attention sublayer at every VLM transformer block provides direct access to the raw 256-primitive Gaussian field during DA-CoT generation. A 300M-parameter flow-matching action expert with mixture-of-experts feedforward sublayers decodes 7-DoF delta action chunks via conditional ODE integration, conditioned on both VLM hidden states and DA-CoT outputs through dual cross-attention. Trained with composite $\mathcal{L}_\mathrm{flow} + \mathcal{L}_\mathrm{CoT} + \mathcal{L}_\mathrm{depth}$ across three progressive stages, GST-VLA achieves 96.4% on LIBERO (+2.0%), and 80.2% on SimplerEnv (+5.4%). Ablations isolate the contribution of each GST component, each DA-CoT thought, and each training stage, confirming independent and synergistic gains concentrated on precision demanding tasks.

3D Visual Grounding (3DVG) localizes objects from natural-language descriptions in 3D scenes and is fundamental to embodied AI applications. Although foundation models enable open-vocabulary reasoning, they typically rely on pre-generated candidates, creating two sequential bottlenecks. The candidate bottleneck occurs when dataset-specific 3D proposal models miss, fragment, or incorrectly group targets under distribution shifts, excluding them from VLM reasoning. The evidence bottleneck stems from incomplete visual evidence: global renderings preserve spatial context but obscure object details, whereas candidate-centric views capture local appearance but lack global context. To address these bottlenecks, we propose UniGround, a zero-shot 3DVG framework that addresses both bottlenecks through Global Candidate Filtering and Contextual Precision Grounding. Global Candidate Filtering constructs topology-consistent, class-agnostic candidates from 3D topology and multi-view semantic cues, without dataset-trained 3D detectors, task-specific proposal supervision, or predefined box and category priors. Contextual Precision Grounding jointly reasons over global spatial context and candidate-centric visual evidence, followed by closed-loop consistency verification for reliable target identification. UniGround achieves 46.1%/34.1% [email protected]/0.5 on ScanRefer and 28.7% [email protected] on the evaluated ARKitScenes subset of EmbodiedScan. Further experiments demonstrate competitive grounding without dataset-specific 3D priors, cross-dataset generalization to unseen indoor scenes, and robustness to real-world reconstruction noise and practical domain shifts.

Localizing objects and parts from natural language in 3D space is essential for robotics, AR, and embodied AI, yet existing methods face a trade-off between the accuracy and geometric consistency of per-scene optimization and the efficiency of feed-forward inference. We present TrianguLang, a feed-forward framework for 3D localization that requires no camera calibration at inference. Unlike prior methods that treat views independently, we introduce Geometry-Aware Semantic Attention (GASA), which utilizes predicted geometry to gate cross-view feature correspondence, suppressing semantically-plausible but geometrically-inconsistent matches without requiring ground-truth poses. Validated on five benchmarks including ScanNet++ and uCO3D, TrianguLang achieves state-of-the-art feed-forward text-guided segmentation and localization, reducing user effort from $O(N)$ clicks to a single text query. The model processes each frame at 1008x1008 resolution in $\sim$57ms ($\sim$18 FPS) without optimization, enabling practical deployment for interactive robotics and AR applications. Code and checkpoints are available at https://cwru-aism.github.io/triangulang/.

The pursuit of spatial intelligence fundamentally relies on access to large-scale, fine-grained 3D data. However, existing approaches predominantly construct spatial understanding benchmarks by generating question-answer (QA) pairs from a limited number of manually annotated datasets, rather than systematically annotating new large-scale 3D scenes from raw web data. As a result, their scalability is severely constrained, and model performance is further hindered by domain gaps inherent in these narrowly curated datasets. In this work, we propose Holi-Spatial, the first fully automated, large-scale, spatially-aware multimodal dataset, constructed from raw video inputs without human intervention, using the proposed data curation pipeline. Holi-Spatial supports multi-level spatial supervision, ranging from geometrically accurate 3D Gaussian Splatting (3DGS) reconstructions with rendered depth maps to object-level and relational semantic annotations, together with corresponding spatial Question-Answer (QA) pairs. Following a principled and systematic pipeline, we further construct Holi-Spatial-4M, the first large-scale, high-quality 3D semantic dataset, containing 12K optimized 3DGS scenes, 1.3M 2D masks, 320K 3D bounding boxes, 320K instance captions, 1.2M 3D grounding instances, and 1.2M spatial QA pairs spanning diverse geometric, relational, and semantic reasoning tasks. Holi-Spatial demonstrates exceptional performance in data curation quality, significantly outperforming existing feed-forward and per-scene optimized methods on datasets such as ScanNet, ScanNet++, and DL3DV. Furthermore, fine-tuning Vision-Language Models (VLMs) on spatial reasoning tasks using this dataset has also led to substantial improvements in model performance.

Generalised 3D Referring Expression Segmentation (3D-GRES) localizes objects in 3D scenes based on natural language, even when descriptions match multiple or zero targets. Existing methods rely solely on sparse point clouds, lacking rich visual semantics for fine-grained descriptions. We propose HCF-RES, a multi-modal framework with two key innovations. First, Hierarchical Visual Semantic Decomposition leverages SAM instance masks to guide CLIP encoding at dual granularities -- pixel-level and instance-level features -- preserving object boundaries during 2D-to-3D projection. Second, Progressive Multi-level Fusion integrates representations through intra-modal collaboration, cross-modal adaptive weighting between 2D semantic and 3D geometric features, and language-guided refinement. HCF-RES achieves state-of-the-art results on both ScanRefer and Multi3DRefer.

Reinforcement Learning with Verifiable Rewards ( RLVR ) has emerged as a transformative paradigm for enhancing the reasoning capabilities of Large Language Models ( LLMs), yet its potential in 3D scene understanding remains under-explored. Existing approaches largely rely on Supervised Fine-Tuning ( SFT), where the token-level cross-entropy loss acts as an indirect proxy for optimization, leading to a misalignment between training objectives and task performances. To bridge this gap, we present Reinforcement Fine-Tuning for Video-based 3D Scene Understanding (3D-RFT ), the first framework to extend RLVR to video-based 3D perception and reasoning. 3D-RFT shifts the paradigm by directly optimizing the model towards evaluation metrics. 3D-RFT first activates 3D-aware Multi-modal Large Language Models ( MLLM s) via SFT, followed by reinforcement fine-tuning using Group Relative Policy Optimization ( GRPO) with strictly verifiable reward functions. We design task-specific reward functions directly from metrics like 3D IoU and F1-Score to provide more effective signals to guide model training. Extensive experiments demonstrate that 3D-RFT-4B achieves state-of-the-art performance on various video-based 3D scene understanding tasks. Notably, 3D-RFT-4B significantly outperforms larger models (e.g., VG LLM-8B) on 3D video detection, 3D visual grounding, and spatial reasoning benchmarks. We further reveal good properties of 3D-RFT such as robust efficacy, and valuable insights into training strategies and data impact. We hope 3D-RFT can serve as a robust and promising paradigm for future development of 3D scene understanding.

Current 3D visual grounding tasks only process sentence level detection or segmentation, which critically fails to leverage the rich compositional contextual reasonings within natural language expressions. To address this challenge, we introduce Detailed 3D Referring Expression Segmentation (3D-DRES), a new task that provides a phrase to 3D instance mapping, aiming at enhancing fine-grained 3D vision language understanding. To support 3D-DRES, we present DetailRefer, a new dataset comprising 54,432 descriptions spanning 11,054 distinct objects. Unlike previous datasets, DetailRefer implements a pioneering phrase-instance annotation paradigm where each referenced noun phrase is explicitly mapped to its corresponding 3D elements. Additionally, we introduce DetailBase, a purposefully streamlined yet effective baseline architecture that supports dual-mode segmentation at both sentence and phrase levels. Our experimental results demonstrate that models trained on DetailRefer not only excel at phrase-level segmentation but also show surprising improvements on traditional 3D-RES benchmarks.

Humans commonly identify 3D object affordance through observed interactions in images or videos, and once formed, such knowledge can be generically generalized to novel objects. Inspired by this principle, we advocate for a novel framework that leverages emerging multimodal large language models (MLLMs) for interaction intention-driven 3D affordance grounding, namely HAMMER. Instead of generating explicit object attribute descriptions or relying on off-the-shelf 2D segmenters, we alternatively aggregate the interaction intention depicted in the image into a contact-aware embedding and guide the model to infer textual affordance labels, ensuring it thoroughly excavates object semantics and contextual cues. We further devise a hierarchical cross-modal integration mechanism to fully exploit the complementary information from the MLLM for 3D representation refinement and introduce a multi-granular geometry lifting module that infuses spatial characteristics into the extracted intention embedding, thus facilitating accurate 3D affordance localization. Extensive experiments on public datasets and our newly constructed corrupted benchmark demonstrate the superiority and robustness of HAMMER compared to existing approaches. All code and weights are publicly available.

Pre-trained general-purpose Vision-Language Models (VLM) hold the potential to enhance intuitive human-machine interactions due to their rich world knowledge and 2D object detection capabilities. However, VLMs for 3D coordinates detection tasks are rare. In this work, we investigate interactive abilities of VLMs by returning 3D object positions given a monocular RGB image from a wrist-mounted camera, natural language input, and robot states. We collected and curated a heterogeneous dataset of more than 100,000 images and finetuned a VLM using QLoRA with a custom regression head. By implementing conditional routing, our model maintains its ability to process general visual queries while adding specialized 3D position estimation capabilities. Our results demonstrate robust predictive performance with a median MAE of 13 mm on the test set and a five-fold improvement over a simpler baseline without finetuning. In about 25% of the cases, predictions are within a range considered acceptable for the robot to interact with objects.

While Multimodal Large Language Models (MLLMs) excel in semantic tasks, they frequently lack the "spatial sense" essential for sophisticated geometric reasoning. Current models typically suffer from exorbitant modality-alignment costs and deficiency in fine-grained structural modeling precision.We introduce SSR, a framework designed for Structured Scene Reasoning that seamlessly integrates 2D and 3D representations via a lightweight alignment mechanism. To minimize training overhead, our framework anchors 3D geometric features to the large language model's pre-aligned 2D visual semantics through cross-modal addition and token interleaving, effectively obviating the necessity for large-scale alignment pre-training. To underpin complex spatial reasoning, we propose a novel scene graph generation pipeline that represents global layouts as a chain of independent local triplets defined by relative coordinates. This is complemented by an incremental generation algorithm, enabling the model to construct "language-model-friendly" structural scaffolds for complex environments. Furthermore, we extend these capabilities to global-scale 3D global grounding task, achieving absolute metric precision across heterogeneous data sources. At a 7B parameter scale, SSR achieves state-of-the-art performance on multiple spatial intelligence benchmarks, notably scoring 73.9 on VSI-Bench. Our approach significantly outperforms much larger models, demonstrating that efficient feature alignment and structured scene reasoning are the cornerstones of authentic spatial intelligence.

3D object affordance grounding aims to identify regions on 3D objects that support human-object interaction (HOI), a capability essential to embodied visual reasoning. However, most existing approaches rely on static visual or textual cues, neglecting that affordances are inherently defined by dynamic actions. As a result, they often struggle to localize the true contact regions involved in real interactions. We take a different perspective. Humans learn how to use objects by observing and imitating actions, not just by examining shapes. Motivated by this intuition, we introduce video-guided 3D affordance grounding, which leverages dynamic interaction sequences to provide functional supervision. To achieve this, we propose VAGNet, a framework that aligns video-derived interaction cues with 3D structure to resolve ambiguities that static cues cannot address. To support this new setting, we introduce PVAD, the first HOI video-3D pairing affordance dataset, providing functional supervision unavailable in prior works. Extensive experiments on PVAD show that VAGNet achieves state-of-the-art performance, significantly outperforming static-based baselines. The code and dataset will be open publicly.

3D large multimodal models (3D LMMs) rely heavily on ego poses for enabling directional question-answering and spatial reasoning. However, most existing point cloud benchmarks contain rich directional queries but lack the corresponding ego poses, making them inherently ill-posed in 3D large multimodal modelling. In this work, we redefine a new and rigorous paradigm that enables direction-aware 3D LMMs by identifying and supplementing ego poses into point cloud benchmarks and transforming the corresponding point cloud data according to the identified ego poses. We enable direction-aware 3D LMMs with two novel designs. The first is PoseRecover, a fully automatic pose recovery pipeline that matches questions with ego poses from RGB-D video extrinsics via object-frustum intersection and visibility check with Z-buffers. The second is PoseAlign that transforms the point cloud data to be aligned with the identified ego poses instead of either injecting ego poses into textual prompts or introducing pose-encoded features in the projection layers. Extensive experiments show that our designs yield consistent improvements across multiple 3D LMM backbones such as LL3DA, LL3DA-SONATA, Chat-Scene, and 3D-LLAVA, improving ScanRefer mIoU by 30.0% and Scan2Cap LLM-as-judge accuracy by 11.7%. In addition, our approach is simple, generic, and training-efficient, requiring only instruction tuning while establishing a strong baseline for direction-aware 3D-LMMs.

Vision-Language Models (VLMs) have demonstrated strong reasoning capabilities over images and videos, yet their application to embodied scene understanding often constrained by the fixed viewpoints stored in episodic RGB-D memories. These observations may fail to capture query-relevant evidence due to occlusions, object truncation, restricted fields of view, or suboptimal view composition. We present SplatReasoner, a framework that introduces novel view synthesis into the VLM reasoning process by leveraging 3D Gaussian Splatting (3DGS). Given a user query about a 3D scene, SplatReasoner retrieves relevant observations and synthesizes query-conditioned viewpoints that reveal the visual evidence needed to answer the query and ground the referred entities in 3D. Experiments show that query-conditioned novel view synthesis improves both embodied reasoning and 3D grounding over fixed-view memory and language-embedded 3DGS baselines.

Recent progress in medical vision-language models (VLMs) has achieved strong performance on image-level text-centric tasks such as report generation and visual question answering (VQA). However, achieving fine-grained visual grounding and volumetric spatial reasoning in 3D medical VLMs remains challenging, particularly when aiming to unify these capabilities within a single, generalizable framework. To address this challenge, we proposed MedVL-SAM2, a unified 3D medical multimodal model that concurrently supports report generation, VQA, and multi-paradigm segmentation, including semantic, referring, and interactive segmentation. MedVL-SAM2 integrates image-level reasoning and pixel-level perception through a cohesive architecture tailored for 3D medical imaging, and incorporates a SAM2-based volumetric segmentation module to enable precise multi-granular spatial reasoning. The model is trained in a multi-stage pipeline: it is first pre-trained on a large-scale corpus of 3D CT image-text pairs to align volumetric visual features with radiology-language embeddings. It is then jointly optimized with both language-understanding and segmentation objectives using a comprehensive 3D CT segmentation dataset. This joint training enables flexible interaction via language, point, or box prompts, thereby unifying high-level visual reasoning with spatially precise localization. Our unified architecture delivers state-of-the-art performance across report generation, VQA, and multiple 3D segmentation tasks. Extensive analyses further show that the model provides reliable 3D visual grounding, controllable interactive segmentation, and robust cross-modal reasoning, demonstrating that high-level semantic reasoning and precise 3D localization can be jointly achieved within a unified 3D medical VLM.

3D visual grounding aims to locate objects based on natural language descriptions in 3D scenes. Existing supervised methods are limited by generalization and recent zero-shot methods typically rely on a predefined Object Lookup Table (OLT) to query Visual Language Models (VLMs) for reasoning about object locations via a single step grounding, which limits the applications in scenarios with undefined targets and complex queries. To address these problems, we present OpenGround, a novel zero-shot framework for open-world 3D visual grounding that remains compatible with recent zero-shot methods. OpenGround integrates Task-Chain Planning to decompose a query into a plan of context-to-target sub-goals for progressive grounding, and Context-Guided Perception to perceive novel objects online under context guidance from the task chain. We also propose a new dataset named OpenTarget, which contains over 7000 object-description pairs to mimic open-world evaluation. Extensive experiments demonstrate that OpenGround achieves competitive performance on Nr3D, state-of-the-art on ScanRefer, and delivers a substantial 17.6% improvement on OpenTarget. Project Page at https://why-102.github.io/openground.io/.

3D Visual Grounding (3DVG) is a critical bridge from vision-language perception to robotics, requiring both language understanding and 3D scene reasoning. Traditional supervised models leverage explicit 3D geometry but exhibit limited generalization, owing to the scarcity of 3D vision-language datasets and the limited reasoning capabilities compared to modern vision-language models (VLMs). We propose a generalizable 3DVG framework, PanoGrounder, that couples multi-modal panoramic representation with pretrained 2D VLMs for strong vision-language reasoning. Panoramic renderings, augmented with 3D semantic and geometric features, serve as an intermediate representation between 2D and 3D, and offer two major benefits: (i) they can be directly fed to VLMs with minimal adaptation and (ii) they retain long-range object-to-object relations thanks to their 360-degree field of view. We devise a three-stage pipeline that places a compact set of panoramic viewpoints considering the scene layout and geometry, grounds a text query on each panoramic rendering with a VLM, and fuses per-view predictions into a single 3D bounding box via lifting. Our approach achieves state-of-the-art results on ScanRefer and Nr3D, and demonstrates strong generalization to unseen 3D datasets and text rephrasings.

Spatial tracing, as a fundamental embodied interaction ability for robots, is inherently challenging as it requires multi-step metric-grounded reasoning compounded with complex spatial referring and real-world metric measurement. However, existing methods struggle with this compositional task. To this end, we propose RoboTracer, a 3D-aware VLM that first achieves both 3D spatial referring and measuring via a universal spatial encoder and a regression-supervised decoder to enhance scale awareness during supervised fine-tuning (SFT). Moreover, RoboTracer advances multi-step metric-grounded reasoning via reinforcement fine-tuning (RFT) with metric-sensitive process rewards, supervising key intermediate perceptual cues to accurately generate spatial traces. To support SFT and RFT training, we introduce TraceSpatial, a large-scale dataset of 30M QA pairs, spanning outdoor/indoor/tabletop scenes and supporting complex reasoning processes (up to 9 steps). We further present TraceSpatial-Bench, a challenging benchmark filling the gap to evaluate spatial tracing. Experimental results show that RoboTracer surpasses baselines in spatial understanding, measuring, and referring, with an average success rate of 79.1%, and also achieves SOTA performance on TraceSpatial-Bench by a large margin, exceeding Gemini-2.5-Pro by 36% accuracy. Notably, RoboTracer can be integrated with various control policies to execute long-horizon, dynamic tasks across diverse robots (UR5, G1 humanoid) in cluttered real-world scenes. Please see the project page at https://zhoues.github.io/RoboTracer.

Despite recent progress in 3D-LLMs, they remain limited in accurately grounding language to visual and spatial elements in 3D environments. This limitation stems in part from training data that focuses on language reasoning rather than spatial understanding due to scarce 3D resources, leaving inherent grounding biases unresolved. To address this, we propose 3D scene editing as a key mechanism to generate visual counterfactuals that mitigate these biases through fine-grained spatial manipulation, without requiring costly scene reconstruction or large-scale 3D data collection. Furthermore, to make these edits targeted and directly address the specific weaknesses of the model, we introduce DEER-3D, an error-driven framework that diagnoses grounding failures and generates targeted counterfactual training supervision via a structured "Decompose, Diagnose, Edit, and Retrain" loop. Specifically, given a grounding failure, DEER-3D first identifies the predicate-level error (e.g., attribute or spatial relation). It then performs minimal predicate-aligned scene edits, such as recoloring or repositioning, and constructs aligned question-answer pairs that explicitly target the failed predicate, forming targeted counterfactual training examples. We evaluate our editing pipeline across multiple benchmarks for 3D grounding and scene understanding tasks, consistently demonstrating improvements across all grounding datasets through iterative refinement (4-6% gains). DEER-3D underscores the effectiveness of targeted, error-driven scene editing in bridging linguistic reasoning with spatial grounding in 3D LLMs.

3D affordance grounding aims to understand how diverse objects can be manipulated, making it a cornerstone of embodied interaction. However, prior works struggle to generalize to out-of-distribution, open-world scenarios, leaving a critical gap between limited dataset performance and real-world application needs. Inspired by the saying: ``What I can not create, I do not understand'', we find generative models can generate semantically valid HOI images, which indicates inherent encoding of affordance concepts. Building on this insight, we propose DAG, the first innovative diffusion-based 3D affordance grounding framework that extracts general affordance knowledge from text-to-image diffusion models for 3D affordance prediction. Specifically, we extract the affordance priors from a diffusion model to encode HOI priors, and design an affordance block with a multi-source affordance decoder for dense 3D affordance prediction. Extensive experiments show that DAG consistently outperforms state-of-the-art methods and exhibits strong open-world generalization, even in the challenging one-shot setting. The code of our method is released on \textcolor{blue}{*https://github.com/hq-King/DAG*}.

Large language models are increasingly being applied to 3D scene understanding. However, existing 3D-aware LLMs largely operate as black boxes: they produce bounding boxes or textual responses without exposing the reasoning process, and they typically depend on pre-trained 3D detectors for object proposals. We present Scene-R1, a video-grounded framework for 3D scene reasoning that avoids both pre-trained 3D object detectors and dense point-cloud backbones. The framework combines reinforcement-learning-based reasoning with a two-stage grounding pipeline. In the first stage, Scene-R1 performs temporal grounding by identifying the video segments most relevant to an open-ended query. In the second stage, it grounds the target in image space by predicting a 2D bounding box. This prediction is then refined into a precise 3D localization using high-quality candidates from a zero-shot segmentation module, enabling fine-grained geometric reasoning without detector-based proposals. Scene-R1 can also be adapted to 3D visual question answering, generating free-form answers directly from RGB-D video. Our training pipeline uses lightweight 2D supervision projected from sparse scans, avoiding the computational overhead required to train 3D instance segmentation models. Across multiple datasets, Scene-R1 outperforms existing open-vocabulary baselines while also providing interpretable intermediate reasoning steps. These results suggest that reinforcement-learning-based reasoning over RGB-D video can support accurate 3D scene understanding with task-level supervision. The training, inference, and 3D refinement code is available at https://github.com/819346675/Scene-R1.

Affordance grounding aims to localize where to interact with an object, a fundamental capability for embodied agents. Yet progress is bottlenecked by data: manual annotation is prohibitively expensive and confines existing datasets to a narrow set of predefined object and affordance categories. We introduce Affogato, a framework for open-vocabulary affordance grounding centered on Affogato-750K, a large-scale dataset of 750K 3D affordance heatmaps paired with natural language queries. We build it with a fully automated pipeline that orchestrates foundation models to generate them at scale without human labeling. It covers significantly more diverse categories than any existing dataset. For reliable evaluation, we further provide 5K human-verified test pairs. We also present Espresso-3D and Espresso-2D, simple yet effective models with a unified architecture across both modalities. Pretraining on Affogato-750K improves both Espresso and prior methods and yields the largest gains on unseen object and affordance categories, showing that it provides broadly transferable supervision across architectures.