Scene Representation

Scene safety understanding plays a life-or-death role in situational awareness in various critical domains. Traditional methods that rely on learning direct mappings between scenes and safety levels often lack interpretability, limiting their reliability in critical applications. An effective approach to overcoming this challenge lies in interpreting human cognitive processes and equipping machine models with analogous cognitive capabilities. This work explores an effective way of integrating scene safety cognitive process modeling and process supervision. Specifically, we first construct a hierarchical cognitive safety structure, which motivates the development of a novel, high-quality scene safety understanding dataset based on multi-step reasoning with process labels. This dataset serves both as a benchmark and a resource to improve the safety reasoning capabilities of Large Language Models (LLMs), while also enabling a granular analysis of intermediate reasoning steps through information flow and saliency-based techniques. Building upon this foundation, we introduce a modular and flexible process supervision framework that reflects the hierarchical nature of human cognition. This framework leverages LLMs as the core architecture and incorporates Low-Rank Adaptation(LoRA) and Mixture-of-Experts (MoE) strategies to enable specialization and collaboration among expert modules, each tasked with specific sub-processes of the overall reasoning chain. Systematic experimental evaluations and analyses confirm that our framework exhibits superior interpretability and performance characteristics compared to traditional approaches.

Open-vocabulary scene understanding is fundamental for robotics, laying the groundwork for spatial reasoning and object manipulation. While closed-vocabulary 3D instance segmentation heavily leverages 3D shape information, state-of-the-art open-vocabulary methods remain predominantly restricted to 2D image features or image-distilled representations during mask labeling. In this paper, we propose SenseFuse, a label-free fusion method that balances 2D image and 3D shape encoders for robust open-vocabulary 3D instance segmentation, refining only the mask-labeling stage of existing pipelines. We reveal that 2D image and 3D shape encoders exhibit largely disjoint failure patterns and rarely share identical wrong labels, whereas two 2D image encoders frequently repeat the same errors. This distinct behavior makes the 2D and 3D pair inherently complementary. We introduce an adaptive mechanism that selects a scene-level fusion weight to maximize a label-free sensitivity measure, estimated directly from a single scene's unlabeled proposals in milliseconds. SenseFuse improves labeling accuracy in every evaluated setting across ScanNet200, Replica, and ScanNet++, recovering 67-100% (median 93%) of the gain achievable with an oracle weight, and it raises instance AP in 21 of 22 reported settings. Code is available at https://github.com/hanes1207/SenseFuse.

Robot manipulation policies are improving quickly, and real-robot evaluation remains the standard evidence for that progress. It still relies on a human to reset the scene between rollouts, which consumes operator time and leaves the initial state distribution unspecified, so results reproduce poorly. A recent system, AutoEval, automates both reset and scoring, but only for single-step tasks, because a long-horizon rollout can terminate in combinatorially many configurations that no single learned reset policy covers. We present HALTER, a Harness for Autonomous Long-horizon Task Evaluation and Reset, which restores the scene by planning over a library of learned atomic reset skills, so demonstration cost scales with the size of that library rather than with the number of terminal states. HALTER builds a spatial scene graph online from point clouds and vision foundation models, and an LLM reasons over this graph to score the rollout, plan the reset, and verify that the reset succeeded, without collecting labeled success images for any task. On four long-horizon tasks on a Franka arm, HALTER restores the scene in 76% of episodes, against 52% for AutoEval and 65% for a motion-planning reset, and it estimates the completed-skill fraction correctly in 90% of episodes, against 76%. Its reset-verification verdict is correct in 91% of episodes, compared with 78% for AutoEval. It also cuts the operator time of an evaluation campaign by 72% relative to manual reset. We further measure compositional generalization on three held-out tasks, where HALTER resets 74.7% of episodes against 1.3% for a per-task reset policy, and we ablate the scene representation and the graph update rate.

Remote sensing recognition is often constrained by scarce observations of rare targets and costly annotations, making realistic synthetic augmentation particularly valuable for few-shot and long-tailed scenarios. Object insertion provides an efficient way to increase target diversity while preserving authentic background scenes, but realistic insertion in overhead imagery requires the generated target to adapt coherently to its surrounding environment. To this end, we propose PDA++, a unified environment-aware object insertion framework organized as Plan, Decouple, and Assimilate. Planning determines scene-compatible poses through an affordance field that combines geometric clearance with structure- and scale-aware cues. Decoupling introduces a pose-conditioned background that provides precise spatial guidance together with target-scene context, allowing the reference object to preserve its identity while adapting to the target observation. This construction also naturally provides pixel-level masks for segmentation augmentation. Assimilation further improves local coherence by aligning multi-scale texture distributions through optimal transport. On the optical benchmark, PDA++ achieves a whole-image FID of 6.28 and improves average few-shot recognition mAP50 by 17.69 points, corresponding to a 28.8% relative gain over the real-data baseline. On SAR imagery, it improves ship detection by 4.10 mAP50 points and remains effective under cross-dataset transfer and amorphous-target insertion. Code is available at https://github.com/lisheyu972/PDA_PLUS.

Interactive control for video generation is moving from coarse prompts toward fine-grained, physically meaningful manipulation of dynamic scenes. Yet existing controllable methods either require the full control schedule before generation starts, or use pixel-space signals that dictate object positions rather than physical dynamics. To address these limitations, we propose PhysStream, an autoregressive model for physics-grounded image-to-video synthesis that incorporates structured scene memory---positional maps and object tracking maps derived online from previously generated frames---and supports fine-grained motion control via sparse velocity-increment signals that encode physical quantities, letting the model learn the underlying dynamics. We train our model in two stages: a bidirectional model is first finetuned with motion-control conditioning, then a causal autoregressive model is trained with additional structured scene memory, further improving physical consistency. PhysStream enables interactive, mid-generation control over multi-object tabletop rigid-body scenes---a capability not supported by prior methods---reducing motion distribution distance (FVMD) by 33% and trajectory error by 12% over the strongest baselines on synthetic benchmarks, and is preferred by human evaluators in over 85% of in-the-wild comparisons. Please check our website for more details: https://czzzzh.github.io/PhysStream

Language-guided robots need persistent scene memories to follow instructions, revisit objects, and resolve references to objects encountered over time. While much of language-guided scene-memory retrieval has emphasized spatial or relational references, many everyday object references specify objects by multiple persistent attributes, such as category, material, size, or surface appearance. We formalize this problem as attribute-compositional retrieval, where a fixed object-centric scene memory is queried with natural language to retrieve the object satisfying the requested attributes. To investigate this capability directly, we introduce a controlled evaluation protocol with fixed scene memories and attribute-defined targets, separating retrieval from perception and annotation ambiguities. We then propose FRAME, which turns language into query-relevant attribute weights, uses learned readouts to estimate per-attribute evidence from object embeddings, and ranks objects by aggregating this evidence according to the query. Across held-out scenes and object assets, FRAME outperforms representative scene-memory retrieval baselines while reducing post-decomposition object scoring to lightweight matrix-vector computation. These results position attribute-compositional retrieval as a complementary scene-memory capability for language-guided robots, showing that persistent object attributes can be exposed as composable evidence for accurate and efficient multi-attribute retrieval.

Long-horizon target navigation requires a robot to sustain task execution across evolving observations, decisions, and physical interactions. This requires three coupled capabilities: maintaining valid scene memory, revising target beliefs under partial observability, and selecting interaction-feasible navigation endpoints. However, the state underlying each capability is only conditionally valid: scene representations become stale when objects move or disappear, unsuccessful searches alter beliefs over target locations, and geometrically convenient endpoints may still be infeasible for manipulation. To address these challenges, we present OmniNav, which formulates long-horizon navigation as continual inference over a factorized task state posterior coupling scene validity, target belief, and interaction feasibility. For representation, OmniNav incrementally constructs an updatable 3D object scene memory, preventing stale scene evidence from propagating to subsequent decisions. For exploration, it introduces an evidence-aware Bayesian belief-revision mechanism that derives dependency-aware region priors from semantic context, incorporates unsuccessful searches as negative evidence, and updates them for posterior-guided frontier selection. For interaction, OmniNav incorporates manipulation reachability and collision constraints into navigation-endpoint selection and propagates execution feedback through hierarchical closed-loop recovery. Extensive experiments demonstrate that OmniNav achieves the highest success rates among the compared methods on semantic ObjectNav and fine-grained instance navigation benchmarks, remains robust to target relocation, and improves real-world pick-and-place success from 53.3% to 71.7% over an adapted open-loop baseline. The project page of OmniNav is available at https://omni-nav.github.io/.

Gaussian Splatting has been effective in inferring scene representations that excel in novel view synthesis. Multiple splats cooperate seamlessly to synthesize the pixels of novel views and are jointly optimized even though they only affect each other indirectly, via pixels they project to in common. We present an approach that enables direct communication among splats to enhance the geometric structures they form in 3D. This is accomplished by Tensor Voting, which was originally designed to infer structures from noisy inputs and has been adapted here to provide supervision during test-time optimization, leading to more accurate scene geometry. We introduce a new class of 3D losses that do not rely on rendering and can be combined with essentially all losses previously reported in the literature. Our 3D losses are especially effective when the input views are sparse and geometric regularization is essential due to limited supervision from the images. Our method is easy to integrate with a diverse set of backbones, and our experiments on the DTU and Tanks-and-Temples datasets demonstrate that TV-SGS improves the geometry of the outputs compared to the backbone, while maintaining or improving rendering quality.

3D Foundation Models (3DFMs) such as VGGT have recently pushed the boundaries of 3D vision by predicting rich unified representations with feed-foward transformers. The scene representations learned by these models enable strong performance on multiple 3D vision tasks. In this paper, we investigate using their internal representations to infer 3D in the scene from new views. Our hypothesis is that in order to solve the task of 3D reconstruction, these models need to learn a representation that includes a large amount of general knowledge about 3D scenes. After showing that it is possible to decode hidden surfaces from internal 3DFM representations, we propose a method, Z3D, that estimates pointmaps in unseen views by doing latent diffusion on 3DFM representation. We show that Z3D can predict realistic depth maps for new views across multiple datasets.

3D Gaussian Splatting has become a de facto scene representation for novel view synthesis, yet robustly learning 3D Gaussian primitives from visual input remains challenging. Standard optimization relies on gradient-based updates, but a common issue is the gradient vanishing phenomenon: a pixel far from a Gaussian primitive often has diminishing gradient magnitudes to influence primitive attributes, resulting in suboptimal scene reconstruction. In this paper, we propose a method to address gradient vanishing with a piecewise truncated gradient formulation that improves the optimization stability and robustness to initializations. We show that our method consistently improves 3D Gaussian Splatting with random and COLMAP initializations while being generalizable across static and dynamic Gaussian Splatting. As a by-product, we also examine the limitations of current benchmarks for dynamic scenes, and introduce a novel dataset for benchmarking dynamic Gaussian Splatting using synthetic 3D scenes. We demonstrate the effectiveness of our method in both static and dynamic settings for the public benchmarks and our proposed dataset.

End-to-end autonomous driving models plan future trajectories from raw sensor input. While earlier driving benchmarks often measured deviation from the human trajectory, current benchmarks such as NAVSIM and Bench2Drive evaluate models with richer simulation-based metrics intended to capture safe and compliant driving. A high benchmark score should reflect that a model can understand the scene in front of it and act accordingly. But how much of that score specifically comes from reacting to the dynamic part of that scene? To probe this, we remove a model's camera input and replace it with memories from prior drives at the same location. The retrieved memories can provide persistent scene information, including road layout and location-conditioned regularities, but not the current traffic state. Surprisingly, memory is nearly sufficient on NAVSIM, reaching or even exceeding the performance of leading end-to-end methods without actually observing the evaluated scene. Our results suggest that a high NAVSIM score does not require a planner to react to the current traffic scene and should be treated with caution. This effect is benchmark-dependent: driving from memory causes substantially larger performance drops on Bench2Drive and RealEngine. We provide our code at https://github.com/boschresearch/MemoryDrivoR .

Feed-forward Gaussian reconstruction has recently emerged as an efficient approach for driving scene reconstruction. However, prevailing LiDAR-based methods preserve the initial correspondence between observed points and Gaussian primitives, treating the initialized primitive set as the final representation. Unlike optimization-based 3DGS, these methods cannot accumulate gradients during training to determine how the scenes representation should be densified. Meanwhile, the shared sparse backbone only fuses observations from different timestamps implicitly, without explicitly aggregating cross-time evidence for individual primitives. In this paper, we present Learning Gaussian Structure (LGS), a framework that enhances both Gaussian structure and primitive attributes. Our key observation is that changes in local gradient responses induced by a prune or add intervention reveal whether the corresponding structural adjustment benefits reconstruction. Based on this observation, our Gaussian Densify Policy learns a Densify Map comprising Prune and Addition Scores from controlled interventions, and directly adjusts the Gaussian structure during inference. We further develop a compact Cross-Time Point Query that explicitly retrieves and aggregates neighboring features from Gaussian primitives at other timestamps for reliable attribute prediction. Extensive experiments on the Waymo Open Dataset and PandaSet demonstrate that LGS consistently outperforms existing methods.

Robots operating in human environments need memories that capture not only what objects exist and where, but also how people use them over time and how individual interactions compose into goal-directed activities. Existing 4D scene graphs preserve object and place histories but omit activity structure, whereas activity representations are either not grounded in persistent 3D scenes or rely on externally provided event boundaries and object associations. We present GESTO (Grounded Event and Spatio-Temporal memOry), a spatio-temporal memory that couples a persistent 4D scene graph with a two-level hierarchy of atomic human--object interactions and goal-driven events. From an RGB-D observation stream, GESTO automatically extracts timestamped interactions, grounds them to persistent scene entities, groups them into events, and uses event context to refine uncertain object associations. A relation-aware tool-calling agent queries the resulting memory for activity-centric spatio-temporal reasoning. We evaluate GESTO on the reproducible text, binary, and time categories of an existing benchmark, together with 40 new Space2Event and Event2Space queries. GESTO achieves scores of 0.71, 0.75, and 0.70 on the standard categories, approaching a method supplied with ground-truth event and object grounding, while substantially outperforming the same reasoning framework when these inputs are removed. It further achieves 0.73 and 0.75 on Space2Event and Event2Space queries. Ablations show that hierarchical event structure and context-aware grounding refinement provide complementary benefits, supporting activity-grounded hierarchical memory for retrospective reasoning in dynamic human environments.

We introduce LEGO for advanced open-vocabulary scene understanding. Beyond basic concept recognition, its core innovation lies in capturing the intrinsic semantic hierarchies within the scene, such as the "flowerpot -> bouquet -> bud -> petal" lineage. While foundation models like SAM can identify multi-granular structures in 2D, their partitions are strictly perspective-bound and lack cross-view consensus. LEGO self-adaptively re-grades volatile multi-view SAM granularities into a unified, 3D-consistent hierarchy. This provides precise supervision for the structurally coherent, multi-level segmentation of 3D scenes. By grounding these segments with CLIP embeddings, LEGO recovers open-vocabulary semantic logic across hierarchical levels. Furthermore, by incorporating spatial relationships, we elevate these segments into level-wise language scene graphs, effectively empowering Large Language Models to perform complex, context-aware spatial reasoning and precise visual grounding. Experimental results demonstrate that LEGO establishes new state-of-the-art performance across both promptable and open-vocabulary 3D segmentation benchmarks, exhibiting advanced hierarchical scene decomposition and context-aware spatial reasoning.

Feed-forward 3D Gaussian Splatting (3DGS) enables efficient and generalizable 3D reconstruction, but current feed-forward 3DGS methods for scene understanding remain largely category-oriented. In contrast, instance-aware 3DGS methods typically rely on per-scene optimization and often decouple reconstruction from instance and semantic learning, limiting reciprocal interactions among them. We present InstanceSplat, a unified feed-forward 3DGS framework for generalizable 3D reconstruction and instance-aware scene understanding from pose-free multi-view images. In a single forward pass, InstanceSplat constructs an instance-aware Gaussian representation that jointly encodes appearance, geometry, instance identity, and language-aligned semantics. Shared 3D Gaussians ground instance identities across views, producing renderable and cross-view-consistent instance features. To allow reconstruction and scene understanding to benefit from each other, we further design an instance-centric learning strategy that connects reconstruction, instance learning, and semantic learning through shared instance structure. Specifically, instance cues guide reconstruction, language-aligned semantics strengthen the discrimination of confusing same-category instances, and instance regions aggregate semantic evidence into coherent object-level predictions. Experiments on novel-view synthesis, instance segmentation, and open-vocabulary semantic understanding under varying input-view settings and on an unseen dataset demonstrate state-of-the-art performance, practical efficiency, and strong generalization.

Omni-domain infrared small target (IRST) detection is crucial for infrared surveillance, yet remains challenging due to heterogeneous imaging domains and inconsistent target characteristics. Previous deep learning-based methods have been developed for visual-only paradigms and achieved promising performance on domain-specific tasks. However, existing methods follow the task-specific supervised learning paradigm. This paradigm simplifies the full-scene infrared observations to sparse target supervision, discarding the semantics that remain invariant across heterogeneous domains. Consequently, detection performance suffers substantially under domain shifts. To handle this issue, we introduce \textbf{``understand before detect''}, a paradigm that formulates omni-domain IRST detection as an understanding-driven process, where holistic infrared target understanding precedes precise detection. Building on this paradigm, we propose \textbf{JinSight}, which first develops holistic IRST understanding through language supervision and then transfers the learned cross-domain representations to precise small-target detection. By grounding infrared representations in language semantics, JinSight enables a single model to generalize across heterogeneous infrared domains. We then introduce Latent Semantic Interaction (LSI), which exchanges language-aligned global semantics with fine-grained spatial features in a compact low-rank space. To address the lack of multimodal omni-domain IRST benchmarks, we build \textbf{OmniIRST-VL}, the first large-scale, highly diverse vision--language dataset for omni-domain IRST detection. It comprises over 39k annotations across six complementary instruction tasks covering both scene-level understanding and target-centric reasoning.

Real-to-sim (R2S) depends on scene representations that render observations along robot ego trajectories, yet dense multi-view capture limits per-environment real-image capture-count efficiency, and sparse human capture can leave behavior-scoped robot views under-supported. Camera-controlled synthesis can fill missing views, but its use in R2S requires behavior-admissible queries and capture-anchored structural conditioning. We present R2S-EGO, which couples a simulator-derived robot proxy that represents the behavior-scoped executable query domain with a capture-anchored geometry proxy that supplies scene-specific structural conditions. Within this domain, fixed- budget selection targets current support deficits for which geometry support is available. The generated observations are assimilated as pseudo-observations to refine the visual asset, while real captures remain anchors. The fused geometry proxy also supplies the scene collision surface, which is refreshed between rounds. Together, these updates refine the existing simulation scene while its robot dynamics and control stack stay fixed. Across 48 frozen Unitree G1 ego views in three Replica scenes, six-view R2S-EGO reaches 19.062 dB PSNR, compared with 14.226 dB for the strongest reported R2S baseline. Across five paired policy-training seeds, R2S-EGO achieves 82.5% +/- 6.8% real-G1 sitting success, compared with 10.0% +/- 10.5% for GaussGym.

End-to-end driving requires a coherent understanding of future scenes, yet existing methods model these scenes using task-specific heads and output formats, with limited scalability. Can video generation instead provide a shared predictor? We introduce SUV, a unified end-to-end driving framework that casts future Scene Understanding as Video generation using a pretrained video foundation model. SUV models future appearance, semantics, relative depth, and instance-level dynamics as video streams with a shared video expert, without stream-specific visual prediction heads. Through joint video-action attention, the action expert attends to the latent representations of all future streams and generates the ego trajectory. Experiments show that SUV directly predicts all four future streams, while controlled ablations show that structured future supervision and direct future-stream access yield higher trajectory planning scores. With only a single front camera and no candidate-trajectory selection, SUV outperforms a broad set of recent state-of-the-art methods on both NAVSIM-v2 splits, achieving 91.0 EPDMS on navtest and 36.9 on navhard. On the long-tail WOD-E2E benchmark, SUV achieves a competitive RFS of 7.94.

Despite recent advances in general-purpose robotic manipulation, real-world multi-object clutter remains challenging to handle for today's prevalent approaches. The problem scales in complexity due to more objects and collisions, more unpredictable contact physics, distractors, and task ambiguity. Bridging this gap to real-world deployment requires effective scene abstractions; yet today, producing such abstractions requires extensive task-specific manual engineering, which does not scale. These abstractions are costly to generate and difficult to adjust or fine-tune. We instead propose a plug-and-play fix to automatically generate scene-specific, task-specific, adaptively updating abstractions on top of existing planning and control stacks. LLM-guided Environment Simplification (LENS) produces a de-cluttered abstracted scene representation by merging (e.g., stacked objects) or pruning (e.g., distant objects) scene entities in a closed loop in response to task progress. These dynamic, task-relevant abstractions are versatile and easy to use. In our experiments, we show that LENS improves classical planning, model-based control, and a vision-language-action model, across a diverse set of highly cluttered manipulation scenes. Project website: https://lens-2026.github.io/.

High-level planning for autonomous driving is a knowledge-intensive engineering decision task that requires accurate scene understanding, timely inference, and internally consistent action selection. Vision-language models (VLMs) can make intermediate reasoning explicit, but their use in deployed planners is constrained by costly structured supervision, unnecessary reasoning in routine scenes, and possible inconsistencies between generated rationales and driving actions. We present a cognitive dual-process planning framework that represents planning-relevant scene knowledge in a machine-parsable structured chain-of-thought (S-CoT) schema. An automated data engine integrates perception foundation models, critical-path filtering, and an expert VLM to generate S-CoT supervision without manual annotation of individual rationales. A lightweight visual Arbiter estimates scene complexity from multilevel vision-encoder features before language decoding and routes each input to either fast meta-action prediction or slow structured reasoning. For slow-path outputs, a deterministic rule-based validator checks whether the parsed S-CoT fields are consistent with the final meta-action and provides verifiable rewards for Group Relative Policy Optimization (GRPO). In a 195-scene manual audit, the generated annotations achieve 91.8% CoT accuracy and a 98.5% Logical Consistency Score (LCS). On 574 manually verified NAVSIM test samples, the planner achieves 80.14% planning accuracy and 97.20% LCS while reducing average latency by 17.39% relative to applying slow reasoning to every scene. Evaluation on external long-tail subsets further identifies conditions under which routing and planning performance degrade. Together, these results show how explicit scene knowledge can be operationalized through adaptive reasoning and rule-based verification to support high-level VLM planning decisions.

Recent advancements in 3D Gaussian Splatting (3DGS) have enabled language-guided scene understanding. However, existing Referring 3D Gaussian Splatting (R3DGS) methods are fundamentally restricted to single-target queries. To reflect the ambiguity of real-world instructions, we introduce the Generalized Referring 3D Gaussian Splatting Segmentation (GR3DGS) task, which requires dynamically segmenting an arbitrary number of targets (0, 1, or $N$). To facilitate comprehensive evaluation of this new task, we construct two new benchmarks: GR-LERF and GR-ScanNet. Crucially, existing R3DGS paradigms exhibit fundamental technical bottlenecks that severely limit their performance on the GR3DGS task: they lack intrinsic 3D point-level understanding by operating merely on 2D rendered pixels, and they incur prohibitive computational overhead by requiring per-scene optimization to embed heavy semantic features. To dismantle these bottlenecks, we propose ZeroSplat, a novel training-free and zero-feature framework. ZeroSplat lifts 2D Vision-Language Model (VLM) priors into 3D space through robust multi-view geometric constraints. This strategy enables intrinsic point-level understanding without incurring any additional feature storage. Extensive experiments demonstrate that ZeroSplat significantly outperforms state-of-the-art methods across generalized and single-target scenarios while maintaining exceptional efficiency. Project Page: https://inkmind-ai.github.io/ZeroSplat

Large language models bring instruction following and scene reasoning to end-to-end driving, but their inference latency collides with the control rate a vehicle requires. Existing closed-loop agents hide this gap by invoking the model on alternate simulation ticks and replaying the previous command in between, so half of all control outputs ignore the newest observations. We present a fast-slow architecture that removes this compromise. A frozen 7B vision-language backbone acts as the slow system, digesting navigation instructions and visual history at low frequency while exposing its per-layer key-value cache as a standing representation of the scene. A lightweight action expert acts as the fast system, attending to this cache and to the current camera frame at every simulation tick to regress waypoints in a single forward pass. Since the cache lags behind the world at deployment, we train the expert under randomized staleness, aligning training with asynchronous execution. On LangAuto-Short routes in CARLA, our system produces fresh control at every 50 ms simulation tick and lifts route completion from 37.0 to 94.0 over the frame-skipping baseline. A frame-skip ablation with the same expert separates the two factors at work: the expert raises the driving score on its own, while per-tick freshness raises completion from 82.1 to 94.0 and cuts red-light violations by a third. Trained on a single town, the expert transfers zero-shot to two unseen towns, holding 84-94% route completion where the baseline reaches 31-41%. It reduces open-loop waypoint error by nearly a factor of four compared to the backbone's own action head, at a per-tick model cost of 32 ms that is independent of history length on a single consumer GPU.

Deploying object-centric models for real-world scene understanding typically requires complex pipelines to achieve both robust scene decomposition and high-fidelity generation. Recent diffusion-based approaches have improved visual quality, but they almost universally rely on heavy, pretrained generative priors (e.g., Stable Diffusion) and external VAE latent spaces. In this paper, we propose Slot-RAE, a much simpler, fully integrated framework that operates directly within the continuous semantic feature space of visual foundation models (e.g., DINOv3). Slot-RAE employs a feature-space diffusion process using a Diffusion Transformer (DiT) decoder and a Representation Alignment (REPA) head. Unlike existing diffusion-based objectcentric methods that rely heavily on subsidized text-toimage priors, the generative core of Slot-RAE (Slot Attention and the DiT) is trained from scratch within the frozen VFM feature space. This eliminates the need for VAE bottlenecks and task-agnostic generative pre-training. Experiments on the COCO dataset demonstrate that despite its architectural simplicity, Slot-RAE achieves state-of-the-art results. It delivers comparable unsupervised object discovery, higher-fidelity image reconstruction, and robust zero-shot compositionality, all while being significantly faster and more computationally efficient than existing object-centric latent diffusion models.

Slot attention is a powerful framework for object-centric learning, decomposing visual scenes into latent slots through iterative competitive attention. However, existing methods share two critical limitations: they decompose scenes into a flat set of slots at a single granularity, and this decomposition is based on appearance rather than semantics. Yet humans understand scenes through semantic hierarchies: separating foreground from background, recognizing object categories, and identifying individual instances. Crucially, such semantic hierarchies cannot emerge without supervision, because category names are human constructs, not visual patterns. We propose Hierarchical Slot Attention (HSA), which learns multi-granularity semantic scene decomposition from a single model. HSA decomposes scenes at three levels: holistic (foreground/background), semantic (object categories), and panoptic (individual instances). Using only 10% labeled data, combined with hierarchical alignment loss, HSA learns all three levels jointly. We further introduce grouping purity and containment to measure whether the hierarchy is encoded in representation space, not just output masks. Experiments on COCO and PASCAL VOC demonstrate that HSA outperforms the strongest flat baseline by up to \textbf{$+$41.5} ARI at holistic, \textbf{$+$14.6} at semantic, and \textbf{$+$10.4} at panoptic level on COCO, with even larger gains on Pascal VOC, while requiring a single model instead of three. Code will be made available upon acceptance.

Open-vocabulary 3D scene understanding aims to segment 3D scenes beyond predefined categories by transferring semantic knowledge from vision-language models. Existing methods have advanced this task by lifting language-aligned 2D features into 3D, yet they often rely on context-independent semantic representations, leaving object relationships underexplored for contextual refinement. We propose RelGraphOV, a relationship-aware framework that uses 3D scene graphs to enhance open-vocabulary 3D understanding. Our method constructs relational scene graphs from multi-view observations by leveraging vision-language reasoning to infer object relationships and prune geometrically implausible connections, without manual relationship annotations. To aggregate relational context while avoiding feature interference, we introduce an Adaptive Gated Dual-Stream Contextual GAT that separates dense geometric features and semantic CLIP embeddings, performs edge-guided message passing, and adaptively fuses complementary semantics. A hierarchical contrastive objective further promotes instance-level consistency and category-level discrimination. Experiments on ScanNetV2, ScanNet200, ScanNet$++$, and Replica demonstrate strong performance and generalization ability. Project Page: https://cxavireh.github.io/relgraphov-projectpage

Gaze target estimation, the task of predicting where a person is looking in a scene, is crucial to understanding human attention and intent. It is a challenging task that combines high-level understanding of global scene semantics and precise spatial reasoning using human appearance (e.g. pose, eye orientation). As a result, human-level performance remains elusive for existing models, limiting their practical application. To this end, we propose PaGE (Practical Gaze Estimator), a gaze estimation model that explicitly models the complex interaction between scene and head features. Using a PaGE model with a large ViT-H+ backbone as the teacher, we further distill student models with lighter backbones on a much larger and more diverse unlabeled dataset. The architectural improvements and novel training recipe allow PaGE to achieve state-of-the-art performance on several gaze estimation tasks, outperforming humans in 7 out of 9 metrics while reducing the human-AI gap by at least 60% in the remaining 2. The distilled student models retain most of the teacher's performance while being lightweight enough for practical deployment on robots and consumer devices. The code and model checkpoints are available at our project page.

Safe motion planning in advanced driver-assistance systems and autonomous vehicles requires an accurate understanding of how the surrounding traffic scene is likely to evolve. However, many existing lane-change prediction methods remain centered on a single target vehicle, while multi-agent forecasting approaches often describe scene evolution only through future positions and provide limited explicit information about the maneuver associated with each vehicle. This study proposes a dynamic scene graph attention framework that predicts the lane-change intention and future trajectory of every relevant vehicle within a local traffic scene. The scene is represented as a time-varying interaction graph in which vehicles are modeled as nodes and their spatial and kinematic relationships are encoded through explicit edge features. Temporal graph-attention message passing captures evolving inter-vehicle dependencies and pre-maneuver cues, while an intention-guided decoder links each predicted maneuver to its corresponding future motion. A scene-level consistency objective further encourages compatible multi-vehicle futures. Experiments on the NGSIM I-80, NGSIM US-101, and highD datasets demonstrate consistent improvements over competing baselines. DSiGAT achieves intention prediction accuracies of 90.12% and 90.97% on NGSIM I-80 and US-101, respectively, and reduces trajectory RMSE by up to 52.94% relative to the strongest baseline. It also produces lower inter-agent collision rates and joint displacement errors, indicating more coherent scene-level predictions. Ablation, sensitivity, robustness, and qualitative analyses further validate the contribution of the proposed components and the effectiveness of the scene-focused formulation.

Reliable instance-level scene understanding is a fundamental prerequisite for object-level interactions and high-fidelity 3D representations. While current methods often leverage 2D foundation segmentation models to obtain these priors, their 2D-centric design typically yields fragmented masks and inconsistent predictions across different views. To address these issues, we propose a novel framework that produces consistent 2D instance masks to guide the optimization of 3D Gaussian Splatting (3DGS) feature fields. Our framework consists of three main stages. (1) Multi-Cue Extraction that generates synergistic semantic, geometric, and structural priors from input images. (2) Multi-Cue-Guided Mask Merging process that consolidates fragmented masks using a composite merge score derived from semantic, depth, and edge cues. (3) Cross-View Mask Matching that establishes globally consistent identity assignments across all viewpoints. By transforming viewpoint-specific segments into coherent 3D primitives, our approach enables stable 3D instance segmentation and effective downstream editing tasks. Experiments demonstrate that our method significantly improves cross-view consistency and segmentation stability over existing baselines while maintaining high-fidelity photometric reconstruction.

Comprehensive 3D scene understanding from sparse, unposed images requires a model to recover renderable geometry, open-vocabulary semantics, and free/occupied 3D space without relying on external camera calibration. Recent feed-forward Gaussian methods improve pose-free reconstruction and semantic rendering, but their Gaussian primitives are mainly optimized through image-space objectives and remain weakly constrained in unobserved regions. We propose \textit{COVScene}, a pose-free semantic Gaussian framework that couples renderable Gaussian primitives with a dense semantic occupancy field through differentiable volumetric lifting. Instead of converting Gaussians to voxels only at evaluation time, COVScene lifts the predicted semantic Gaussians inside the training computation graph, so volumetric regularization provides gradients to Gaussian opacity, geometry, and semantic features. The framework combines a semantic-aware Geometry Transformer, multi-task Gaussian decoding, geometric foundation distillation, and occupancy entropy regularization to support novel view synthesis, open-vocabulary semantic querying, and semantic occupancy prediction within a single representation. Experiments on ScanNet and ScanNet++ show that COVScene maintains competitive rendering quality, improves open-vocabulary segmentation, and achieves stronger semantic occupancy prediction than the self-supervised baseline without direct voxel-level supervision.

3D Gaussian Splatting (3DGS) has emerged at the forefront of 3D scene reconstruction. Extending 3DGS with language-driven, open-vocabulary understanding has gained significant attention for real-world applications such as embodied AI. Recent methods achieve this by learning an instance feature attribute and assigning semantics by distilling high-dimensional Contrastive Language-Image Pretraining (CLIP) features directly into the scene representation. However, the instance grouping mechanisms of these methods either require a predefined number of instances or suffer from noise in their bottom-up grouping strategies. Furthermore, the reliance on CLIP restricts semantic understanding to simple noun phrases, preventing complex spatial reasoning and referential expression grounding. We present GaussDet, a method that circumvents the need for dense CLIP features by leveraging discrete, open-vocabulary 2D object detectors with referring expression capabilities. We learn instance features for individual Gaussians to decompose the scene into 3D instance groups. By rendering these groups and aggregating semantic votes from multi-view 2D detections, we generate a robust View-Aggregated Semantic Label Distribution (VASD) for each 3D instance. This view-aggregation strategy acts as a strong regularizer, attenuating spurious labels caused by low-quality instance grouping. Our approach enables a straightforward, zero-shot extension from simple language queries to complex referential grounding. Extensive evaluations across two key tasks -- open-vocabulary segmentation (LeRF-OVS, ScanNet) and referring expression grounding (Ref-LeRF) -- demonstrate that GaussDet achieves consistent improvements over existing methods. Most notably, we achieve a substantial 16.7% mIoU improvement in referential grounding within a strict zero-shot setting.

A 3D scene is understood through its objects, not the primitives that compose them. Yet feed-forward reconstruction methods output dense, unstructured sets of points or Gaussians, leaving object-level structure to be recovered after the fact. We propose a feed-forward framework that decomposes a scene into instance-structured 3D token groups directly from unposed multi-view images -- compact object-centric units from which reconstruction, segmentation, and manipulation all follow. Each token group pairs an instance token capturing entity-level identity with anchor tokens that encode local geometry and appearance, which are decoded into a set of 3D Gaussians. This two-level factorization decouples object identity from local appearance, making object instances a native interface of the representation rather than a derived product. The token groups are learned through differentiable rendering with joint reconstruction and segmentation supervision, requiring no 3D annotations. Our feed-forward model surpasses per-scene optimization baselines in class-agnostic instance segmentation while remaining competitive in novel view synthesis. Beyond these metrics, the same token groups directly unlock instance-level scene editing -- removing, translating, or inserting objects by operating on their groups -- as well as efficient open-vocabulary 3D instance retrieval, where retrieval complexity scales with the number of instances rather than primitives.

Reconstructing dynamic human--object interaction scenes from monocular video is difficult because the human, manipulated object, and background obey different motion models while sharing the same pixels. Existing dynamic radiance-field and Gaussian-splatting methods often entangle these components, causing object motion to leak into the human or static scene, and monocular human reconstruction remains underconstrained in regions that are rarely observed. We present CoGS, a compositional Gaussian-splatting framework for monocular human--object scene reconstruction. CoGS decomposes the video into three coordinated branches: an articulated human initialized from a complete canonical prior, a rigid object field driven by an estimated object trajectory, and a static scene field regularized by weak scene-only planar primitives when available. A six-stage optimization schedule first stabilizes the human and object independently, then fuses them with the scene under full-image supervision, visibility-aware human anchoring, object silhouette and motion constraints, and delayed scene regularization. This design keeps each component responsible for its own geometry and motion while allowing photometric evidence to correct the final composite. Experiments on HOSNeRF and NeuMan show that CoGS improves both human--object interaction reconstruction and in-the-wild human--scene rendering, achieving stronger fidelity and perceptual quality across full-frame and human-focused evaluations. Code will be released upon publication.

Predicting how a scene will evolve after a desired 3D transformation from images is a central goal in vision, graphics, and robotics. Yet unlike ideal simulators with full access to 3D geometry and dynamics, real world systems must rely on perceptual inputs and local actions that are inherently partial and incomplete. In this work, we present P3Sim, a physical world modeling system that simulates future scene states under both partial observations and incomplete 3D transformation signals. P3Sim is composed of three interacting components: a learned physical world model, a geometric conditioning module, and a persistent scene memory. The world model interprets perception as probabilistic inference over multimodal scene variables, providing predictions of the distributions of any scene variable conditioned on any combination of others. The geometric conditioning module provides a partial 3D transform signal for conditioning the world model at inference time. The persistent scene memory integrates predictions over time, enabling online updates and consistency under uncertainty. By combining learned inference with explicit geometric structure, P3Sim balances data-driven flexibility with built-in inductive bias. This design yields a flexible perceptual simulator that generalizes across diverse 3D transformation tasks, such as novel view synthesis, object manipulation, and dynamic scene prediction, advancing toward general purpose 3D scene understanding and transformation.

Embodied reasoning and exploration are increasingly considered crucial abilities for robots operating in complex and unfamiliar environments. To accomplish tasks in such settings, an agent must identify and acquire the information necessary for the task through exploration. We propose ObsGraph, an observation-centric hierarchical scene graph that unifies scene representation, retrieval, and exploration. It retains visual evidence and organizes it into room-view-object layers: rooms provide coarse semantic anchors, views preserve contextual object covisibility, and objects store fine-grained details. On top of this representation, we perform coarse-to-fine hierarchical retrieval under a bounded budget, and crucially use retrieval outcomes to structure the exploration candidate space--activating room-level exploration, view refinement, or frontier exploration--thereby tightly coupling representation, retrieval, and adaptive multi-scale exploration. Experiments across embodied reasoning and exploration benchmarks demonstrate improved success and efficiency, highlighting the benefits of structured scene representation and more targeted information gathering driven by identified evidence gaps.

Pedestrian trajectory prediction requires modeling temporal dynamics, multimodal cues, and social interactions in crowded environments. Existing methods often address these factors separately or entangle them in costly attention blocks, limiting scalability, flexibility, and interpretability. We propose a three-step hierarchical Transformer that explicitly separates temporal encoding, multimodal fusion, and scene-level interaction reasoning. Lightweight GRU summaries enable efficient cross-modal attention, while social attention over time--agent tokens captures inter-pedestrian influences at manageable cost. Experiments on JTA, JRDB, and the Pedestrians and Cyclists in Road Traffic dataset show state-of-the-art performance on real-world datasets (JRDB, Urban) and competitive results on JTA. Ablation and qualitative analyses confirm the contribution of each stage and the model's ability to anticipate complex behaviors such as early turning.

Feed-forward Gaussian splatting (FFGS) facilitates real-time novel view synthesis, yet current methods often remain tied to view-dependent predictions. As more input views are added, they may accumulate noisy or redundant evidence instead of converging to a stable scene representation. In this paper, we introduce CanonicalGS, a feed-forward pipeline that maps cluttered multi-view observations into a stable, scene-centric representation. CanonicalGS first extracts view-centric evidence from depth, semantic features, and uncertainty estimates, and then aggregates this evidence in a canonical latent world using uncertainty-aware fusion. By emphasizing reliable observations while suppressing uncertain or redundant ones, CanonicalGS produces representations that scale more effectively for novel view synthesis and transfer to downstream visual perception tasks. Experiments show up to a $2.5$ dB improvement in peak signal-to-noise ratio for synthesizing novel views and an $11\%$ gain in semantic segmentation accuracy.

3D Gaussian Splatting (3DGS) is widely used to capture and render real scenes. Compositing objects from one capture into another has applications in many domains, such as VFX, architecture and interior design, or marketing. However, extracting an object from a source scene and naively pasting it into a target scene will fail to produce realistic results due to the different lighting conditions between the two scenes. To address this problem, we introduce a diffusion model that harmonizes naively composited images with inconsistent lighting. The model is trained with a heterogeneous dataset of image pairs (inconsistent composite input, consistent output), combining synthetic, generated, and real data. Our complete 3D solution allows a user to extract an object from the source scene and composite it into the target scene. From this, the (inconsistent) views of the target scene with the composite object are rendered. Our diffusion model harmonizes each one of these views, which are finally consolidated in a 3DGS representation with a post-optimization step. Our method provides visually compelling results, making object transfer between 3DGS easy to use and significantly improving quality compared to previous methods.

Real-world spatial intelligence requires reasoning over a continuous and evolving 3D world, yet existing VLMs and tool-augmented agents largely remain tied to static, stateless inference from isolated visual observations. We introduce \textbf{\textsc{S-Agent}}, a spatial tool-use agentic paradigm for understanding and reasoning over continuous multi-view images and videos. By formulating spatial reasoning as spatio-temporal evidence accumulation rather than isolated frame-level prediction, \textsc{S-Agent} reshapes spatial perception into scene-centric understanding beyond frame-centric recognition. Specifically, \textsc{S-Agent} casts the VLM as a semantic planner that decides what evidence is needed, while a hierarchy of spatial tools and experts grounds objects in 2D, lifts them into 3D geometric evidence, and aggregates this evidence into high-level spatial knowledge (\textit{e.g.}, counting, measurement, orientation, and relative position). Additionally, a temporal memory mechanism, including Scene Memory for maintaining the evolving scene state and Agent Memory for accumulating reasoning context, enables evidence integration across frames and reasoning steps. Comprehensive experiments on multi-view and video spatial reasoning benchmarks show that \textsc{S-Agent} consistently improves both open-source and closed-source VLMs in a training-free manner. Beyond inference-time augmentation, supervised fine-tuning (SFT) on \textsc{S-Agent}-generated spatial trajectories \textsc{S-300K} yields \textsc{S-Agent-8B}, a compact spatial agent that significantly surpasses similar-scale baselines (e.g., Qwen3-VL-8B) and performs comparably to advanced closed-source models (e.g., GPT-5.4 and Gemini 3).

Multimodal Large Language Models (MLLMs) have demonstrated remarkable success in various Remote Sensing (RS) tasks. However, their ability to comprehend negation remains underexplored, limiting deployment in real-world applications where models must explicitly identify what is false or absent, e.g., emergency responders need to locate non-flooded routes for evacuation. To comprehensively study this limitation, we introduce RS-Neg, the first benchmark to evaluate negation understanding across region-level to scene-level tasks. Specifically, we design an automated data generation pipeline for RS imagery, using LLMs to synthesize diverse negation queries, and introduce a dynamic visual focus module for verification. Our evaluation reveals that advanced RS MLLMs struggle with negation, exhibiting hallucinations and substantial performance degradation. To close this gap, we propose NeFo, a novel test-time learning method that explicitly incorporates the logical role of negation into the model optimization. Remarkably, using about 5% unlabeled test samples, NeFo significantly improves the negation understanding of models and shows strong generalization to unseen tasks.

Fine-grained aerial object detection, driven by the intrinsic granularity of real-world object categories, is crucial for advanced scene understanding in remote sensing. Existing methods largely inherit the paradigm of coarse-grained object detection, relying solely on single-label supervision and thus struggling to distinguish model-level categories with subtle structural differences. However, for each specific model (e.g., Boeing 787), structured prior knowledge such as attributes and hierarchies offers discriminative semantics across multiple granularities. Motivated by this, we present ExpertDet, a scheme that incorporates expert-informed cues to enhance fine-grained aerial object detection. Specifically, we design Vision-aware Masked Attribute Modeling (VMAM), which aligns attribute semantics with visual structures by reconstructing randomly masked attributes from visual cues, enabling the detector to capture subtle structural distinctions. We further propose Hierarchical Visual Instance Promotion (HierVIP), which builds a visual prototype tree based on hierarchical relations and imposes taxonomy-aware constraints to preserve cross-level semantic continuity while enhancing category discrimination. Moreover, we curate a new fine-grained object detection benchmark for Precise recognition of model-specific Ships and Planes from aerial imagery, PSP, covering 106 ship classes and 30 airplane models, respectively, featuring the most extensive collection of model-specific categories among existing aerial object detection datasets to date. We benchmark state-of-the-art object detection algorithms on the PSP benchmark. Extensive evaluation demonstrates that ExpertDet consistently outperforms other fine-grained competitors across hierarchy levels. The dataset, benchmark, and code are available at https://nnnnerd.github.io/PSP-Benchmark/.

Despite recent advances, Vision Language Models (VLMs) still struggle to grasp the dynamics of the world. We note that the ability to reason about a 4D scene, challenging in itself, is further complicated by two factors. First, VLMs observe motion indirectly via its projection onto 2D images. Second, existing datasets fail to disentangle object and camera motion. To address these challenges, we present a QA generation pipeline that focuses on motion-related scene understanding. We take particular care of the entanglement of camera and object motion by casting tracking in both the traditional way and in a novel, fixed reference system, dubbed True-Motion Tracking, which provides an intuitive description of motion. From this pipeline, we generate a large-scale training dataset of 400K samples, 4DP-QA (4D Perception QA), and a 2.2K-sample benchmark, 4DP-QA-Bench. Training existing models on our dataset yields performance improvements on an external benchmark, validating the effectiveness of our method.

Mining hard, safety-critical scenes from driving logs is bottlenecked by the absence of difficulty labels, and no single proxy, collision risk, trajectory ambiguity, or semantic rarity suffices to find such scenes on its own. We present SceneMiner, a unified, camera-only bird's-eye-view pipeline that emits complementary mining signals from a frozen vision-language backbone in a single forward pass, with no LiDAR or radar: a retrieval embedding for text-prompted scenario search, a multi-label scene-tag distribution, and a continuous physics-based risk score (a motion forecast is a byproduct, not a contribution). Building such a multi-head model exposes our central finding, a failure mode we term cross-task interference: adding or upgrading one head shifts a shared activation stream and degrades weight-frozen sibling heads, so freezing parameters alone is insufficient. Our contribution, identity-preserving multi-task fine-tuning, removes this interference by zero-initializing every new sub-module and freezing every parameter that feeds the shared stream. The mining heads are thereby preserved bit-identically while training only ~102k parameters. The tagging head reaches mAP 0.4614 (micro-F1 0.5557) on 20 scene tags by pooling each scene into 32 visual tokens, and the embedding head supports text-prompted retrieval, validated qualitatively. Code is available at: https://anonymous.4open.science/r/sceneminer_anonymous-64E5

Unauthorized unmanned aerial vehicle (UAV) activity around airports, public venues, and other sensitive sites has made protected-airspace monitoring increasingly important. A practical sensing system must search a wide angular region, find small long-range targets, and return both bearing support and UAV-specific evidence before a restricted perimeter is breached. Existing UAV detection paths often rely on spatially organized evidence, such as body extent, silhouette, or track continuity. At long range, however, these cues become difficult to preserve and verify as the target footprint weakens and its image-plane support shrinks. EventRadar follows a complementary cue: propeller-induced temporal periodicity, which recent event-camera sensing studies have shown can reveal UAV-specific motion after appearance becomes weak. We extend this cue to kilometer-scale active sensing with an event-camera prototype. Scene-Anchored Geometry Evidence (SAGE) fuses scanning events with IMU pose to maintain a bearing-indexed scene memory, separating transient candidate support from persistent background clutter. Comb-guided Harmonic-Group Learned Iterative Shrinkage and Thresholding Algorithm (CHG) then treats each candidate as a weak high-rate timing signal and recovers phase-insensitive harmonic evidence with fixed compute. Compared with related event-camera baselines on 700-1500 m UAV event recordings, EventRadar achieves 0.990 mAP$_{.3}$ and 0.949 F1$_{.3}$, reduces FN$_{.3}$ to 0.009, and shows real-time feasibility in prototype profiling.

Lifelong embodied navigation in dynamic environments requires robots to form persistent scene understanding from fragmentary observations, which remains difficult for existing methods that rely on explicit maps or scene graphs and struggle to generalize beyond structured settings. We propose AllDayNav, a lifelong self-learning navigation framework that implicitly encodes scene dynamics into the billion-scale parameters of a large model via reinforcement learning, powered by a self-evolving multimodal memory that maintains and updates visual keyframes, semantic descriptions, and temporal context while autonomously generating open-vocabulary instructions, image goals, and structured rewards. Experiments in both synthetic and real-world environments across cross-room, cross-episode, and cross-task scenarios show that AllDayNav achieves success rates approaching $100\%$ and consistently surpasses strong map-based, VLM, and RL baselines in path efficiency and robustness, demonstrating implicit, memory-driven reinforcement learning as a scalable alternative to explicit mapping for reliable lifelong navigation.

Perceptual uncertainty is a central challenge for heterogeneous robot teams operating in unstructured outdoor environments, where no single viewpoint affords reliable scene understanding. Perceptual uncertainty, arising from sources such as occlusions, manifests differently across robot viewpoints depending on scene structure. Detecting and resolving sources of perceptual uncertainty requires both scene-based contextual reasoning and capability-aware robot allocation. While vision-language models provide strong semantic priors for both, they are computationally prohibitive for onboard inference and lack calibrated uncertainty quantification. We introduce Co-GLANCE, a real-time onboard perception and decision-making system for uncertainty resolution in heterogeneous robot teams. Co-GLANCE distills the semantic reasoning capabilities of a vision-language model into an end-to-end model for occlusion segmentation and robot allocation, eliminating the need for cloud-based inference. To quantify perceptual uncertainty, Co-GLANCE combines conformal prediction with selective abstention to provide statistically valid coverage guarantees for segmentation, robot allocation, and detection outputs. These calibrated uncertainty estimates directly trigger active perception, dispatching the most appropriate robot to acquire informative viewpoints and resolve uncertainty. Across real-world scenarios, Co-GLANCE outperforms cloud-based vision-language model baselines in occlusion segmentation and robot allocation accuracy by 25% and 36%, respectively, while reducing per-frame inference latency 350x. We also release an air-ground dataset for future research. Code, videos, and dataset available at https://co-glance.github.io/ .

Long-horizon online visual mapping requires continuous camera-motion and scene-geometry estimation under bounded computation. Recent feed-forward 3D reconstruction models provide strong geometric priors, but streaming variants often predict poses in a fixed or historically maintained coordinate system, leading to train--test mismatch, early-anchor attention bias, and accumulated drift. We propose \emph{Anchor3R}, a current-centric streaming 3D reconstruction framework that predicts window-relative poses and local geometry in the current-frame coordinate system. Overlapping predictions form a dense relative-pose graph, supporting online pose updates and loop-aware motion averaging for global reconstruction. Experiments on indoor, outdoor, driving, and RGB-D benchmarks demonstrate improved long-horizon pose accuracy and dense reconstruction quality over existing streaming baselines. Despite being trained only on 48-frame sequences, Anchor3R directly generalizes to streams exceeding 10,000 frames while maintaining bounded GPU memory during online inference. Code is available at https://github.com/polar-explorer/Anchor3R.

Large Language Model (LLM)-based agents increasingly rely on memory to learn from experiences over continual interactions. However, storing experiences as independent, flat units leads to substantial redundancy and retrieval conflicts, as similar episodes repeat overlapping content and subtle scene variations cause retrieved memories to offer contradictory guidance. To address this, we introduce residual experience, positing that newly acquired experience is often an incremental variation of existing knowledge. We propose DeltaMem, a framework that organizes experience memory into two independent residual trees, one storing goal-conditioned task experience as reusable skills and another for scene-level environment knowledge. Each tree uses a root node for generalized base experiences and incremental delta nodes for subsequent variations, allowing related experiences to share a common foundation without duplication. For retrieval, a failure-penalized similarity scan locates the best match, reconstructing the full experience via root-to-match chain composition. An autonomous consolidation mechanism distills high-frequency paths into new root nodes, enabling the trees to self-organize from general heuristics to specialized variants. Experiments across diverse interactive environments show that DeltaMem consistently outperforms existing baselines. To facilitate future research, we release the code at https://github.com/import-myself/DeltaMem.

Reliable mobile manipulation in dynamic indoor environments requires a scene representation that remains geometrically consistent, semantically queryable, and computationally bounded as the environment changes. Existing systems often rely on pre-built maps, static-scene assumptions, or highly accurate camera poses, which can lead to stale or misaligned scene information when target objects are relocated or pose estimates are corrected. This paper presents DREAM, a real-robot mobile manipulation framework that integrates perception, memory, localization, navigation, and manipulation in previously unseen indoor environments without a pre-built map. DREAM constructs an online spatio-semantic voxel memory from RGB-D observations registered by a LiDAR-inertial-visual SLAM backend. It further introduces pose-graph-aware Redundancy-Aware Memory Pruning (RMP) to update historical observations after pose corrections while keeping long-horizon observation history bounded. For target localization and reacquisition, DREAM combines language-conditioned 3D retrieval, open-vocabulary image detection, and multimodal large language model based semantic verification. Real-robot experiments in four dynamic indoor laboratory scenes show that DREAM improves long-horizon task success rates from 40%-60% with DynaMem to 55%-70%, while maintaining a memory footprint of 0.37-0.63 GB and an online memory-update time of 0.43-0.53 s across scenes.

Integrating open-vocabulary semantic information into dynamic 3D scene representations is essential for long-term embodied scene understanding. However, existing methods often suffer from fragile instance association due to incomplete cross-view cues, while their limited ability to handle object-level topological changes restricts long-term robotic task execution. Moreover, current 3D scene understanding methods either rely on simple feature matching without explicit spatial reasoning or assume offline ground-truth 3D geometry. To address these challenges, we present DGSG-Mind, a hybrid instance-aware 3D Gaussian dynamic scene graph system with an embodied reasoning agent. Our system couples a probabilistic voxel grid with explicit 3D Gaussians to enable robust cross-modal instance fusion and incremental semantic mapping. It handles dynamic changes through Gaussian-based visual relocalization and localized masked refinement guided by geometric-semantic consistency. Built on the instance Gaussian map, DGSG-Mind further constructs a hierarchical scene graph and develops the 3D Gaussian Mind, which integrates structural relations, spatial-semantic information, and visually annotated RoI Gaussian renderings for multimodal reasoning. Extensive experiments show that DGSG-Mind achieves the best zero-shot 3DVG performance among methods operating on self-reconstructed maps, while also delivering strong performance in 3D open-vocabulary semantic segmentation and scene reconstruction. We further deploy DGSG-Mind on real-world robots to demonstrate its target-oriented reasoning and dynamic update capabilities. The project page of DGSG-Mind is available at https://icr-lab.github.io/DGSG-Mind

Despite their popularity and success, Multimodal Large Language Models (MLLMs) often struggle to interpret images accurately, which limits their reasoning capability in complex scenarios (e.g., high object density and complex background clutter). Prior work mainly addresses this limitation by incorporating explicit visual cues like bounding boxes that require extra annotations. In addition, the resulting low-resolution crops often miss fine-grained details that MLLMs require for accurate reasoning. Therefore, we propose Mags-RL, an Agentic Reinforcement Learning (RL) framework that equips MLLMs with an external super-resolution "magnifying glass" agent for high-resolution fine-grained inspection. Specifically, the model performs two-round reasoning: in the first round, it generates an initial rationale and autonomously identifies regions of interest without relying on additional annotations; in the second round, it invokes a super-resolution agent to crop and upscale those regions, then revisits and verifies its earlier reasoning to produce the final answer. We also introduce a novel curriculum learning strategy that enables data-efficient RL training, needing as few as only 40 training samples to achieve reasonable performance. Experiments on VSR, TallyQA, and GQA subsets show its superior performance against recent strong competing methods, demonstrating high-quality reasoning with precise visual grounding. Code and weights will be released soon.

Vision-language navigation (VLN) requires an agent to traverse complex 3D environments based on natural language instructions, necessitating a thorough scene understanding. While existing works equip agents with various scene representations to enhance spatial awareness, they often neglect the complex 3D geometry and rich semantics in VLN scenarios, limiting the ability to generalize across diverse and unseen environments. To address these challenges, this work proposes a 3D Gaussian Map that represents the environment as a set of differentiable 3D Gaussians and accordingly develops a navigation strategy for VLN. Specifically, Egocentric Scene Map is constructed online by initializing 3D Gaussians from sparse pseudo-lidar point clouds, providing informative geometric priors for scene understanding. Each Gaussian primitive is further enriched through Open-Set Semantic Grouping operation, which groups 3D Gaussians based on their membership in object instances or stuff categories within the open world, resulting in a unified 3D Gaussian Map. Building on this map, Multi-Level Action Prediction strategy, which combines spatial-semantic cues at multiple granularities, is designed to assist agents in decision-making. Extensive experiments conducted on three public benchmarks (i.e., R2R, R4R, and REVERIE) validate the effectiveness of our method.

Lightweight vision-language models perform competitively on standard benchmarks yet fail systematically in dense-scene reasoning, where multiple objects, attributes, and relations must be jointly grounded and resolved through multi-step inference. Such capability is critical for real-world applications where models must reliably interpret cluttered environments. Yet existing training signals provide no explicit grounding between reasoning steps and the underlying visual entities and relations, leaving lightweight models free to generate fluent but visually unanchored reasoning chains. To address this gap, we first introduce DRBench, a benchmark of 14,573 questions across 2,943 images, organized into five task categories spanning three progressive reasoning layers. Building on DRBench, we propose DRScaffold, a supervised fine-tuning framework that decomposes the supervision target into four causally ordered stages, enforcing grounded reasoning without architectural modification. Experiments on three lightweight VLMs demonstrate substantial gains on DRBench while preserving or improving performance on general-purpose benchmarks. Notably, Qwen2.5-VL-3B trained with DRScaffold surpasses the frozen Qwen2.5-VL-32B on DRBench, demonstrating that structured supervision can substitute for a significant portion of model scale in dense-scene reasoning. Our code and models are available at https://github.com/irene-shi/DRScaffold .

Crucial for autonomous exploration, online 3D occupancy prediction and mapping incrementally constructs dense spatial representations on the fly. However, recent Gaussian-centric methods struggle with structural boundary fidelity and rely heavily on predefined scene-size priors, fundamentally limiting their operational efficiency. In this work, we present VEOcc, a voxel-centric framework formulated as a recursive perception-and-assimilation paradigm. By eliminating the need for initial scale estimation, VEOcc enables highly streamlined, open-ended map expansion. Furthermore, to robustly aggregate noisy temporal observations within the discrete voxel space, we propose a Spatio-Temporal-Aware Online Update Strategy. It integrates Cross-Temporal Logit Aggregation (TLA) for temporal consistency, Reliability-Aware Confidence Modulation (RCM) for spatial uncertainty calibration, and Confidence-Driven Incremental State Update (CSU) for robust global state assimilation. % Extensive experiments on Occ-ScanNet and EmbodiedOcc-ScanNet demonstrate that VEOcc establishes new state-of-the-art performance in both local and embodied settings, providing an accurate and efficient solution for real-world exploration. Extensive experiments on Occ-ScanNet and EmbodiedOcc-ScanNet demonstrate that VEOcc establishes new state-of-the-art performance in both local and embodied settings. Notably, zero-shot evaluations on self-collected video sequences further confirm its robust out-of-distribution generalization capability in completely unseen real-world environments. Ultimately, our framework provides an accurate and highly efficient solution for autonomous exploration. Code and supplementary visualizations are available on our project page: https://wryzju.github.io/VEOcc/.

Mask-free video object insertion has emerged as a challenging task, requiring harmonious integration of reference objects into source videos. However, existing methods struggle when references exhibit severe stylistic domain gaps with the source scene. To overcome this, we propose \textit{\textbf{Smart-Insertion-V}}, an end-to-end \textbf{Dual-Stream} framework that concurrently conducts video insertion and image style transfer. Within this framework, the image stream synchronously guides the video generation process, while a \textbf{Closed-loop Feedback} mechanism is further incorporated to ensure robust insertion. Inevitably, integrating these diverse conditioning signals results in feature entanglement and style leakage. To tackle this issue, we design \textbf{Dual-World-View RoPE} to distinguish different signals via spatial-temporal offsets without incurring heavy training overhead. Furthermore, to facilitate spatial grounding and stylistic adaptation, we introduce a \textbf{Decoupled Guidance Module} that leverages a Vision-Language Model for semantic reasoning while preserving original temporal guidance with native text encoder. To bridge data gap for harmonious reference insertion task, we propose a data curation pipeline and will release an \textbf{open-source dataset}. Experiments demonstrate that our method can insert objects into plausible positions while achieving the most harmonious results.

Spatio-temporal reasoning in vision-language models requires visual representations that preserve physical geometry rather than merely semantic appearance. Recent multimodal models incorporate geometric information through structural branches, 3D-aware supervision, reasoning-stage fusion, or long-horizon memory. While these approaches demonstrate the importance of geometry for spatial intelligence, they typically treat geometric cues as a shared signal across all visual tokens. We note that this overlooks a finer-grained challenge: different visual tokens require different geometric evidence depending on their spatial roles. To address this limitation, we introduce GeoWeaver, a pre-reasoning geometric grounding framework that treats geometry as a representational prerequisite for spatio-temporal reasoning. GeoWeaver constructs a multi-level geometry bank from a frozen geometry encoder and performs token-adaptive geometric evidence allocation, enabling each visual token to retrieve the most relevant geometric abstractions. The selected evidence is incorporated into visual tokens via a residual grounding operation prior to language modeling, yielding geometry-grounded representations for downstream reasoning. Extensive evaluations on spatial reasoning benchmarks demonstrate that GeoWeaver consistently enhances geometry-aware reasoning while retaining general multimodal capabilities. This indicates that geometric information yields the greatest benefit not as a late-fusion auxiliary signal but as a fundamental prerequisite that shapes the representational foundation on which large language models perform reasoning. All source code and models will be released at https://github.com/yahooo-m/GeoWeaver .

Coffee and tea share many properties, yet they evoke strikingly different situations, atmospheres, and affective associations. These situated dimensions of word meaning are real and systematic, but they remain implicit in most computational representations of lexical meaning. We propose Scene Abstraction, a framework for constructing structured representations of the interpretive scenes that words participate in across usage contexts. Each scene consists of a Contextual Scene (Events, Entities, Setting) and an expression-centered Expression Profile (Engaged events, Generalizable properties, Evoked emotions), operationalized through few-shot prompting of a large language model. Our contributions are three-fold: (1) a structured representation framework for situated lexical meaning; (2) COCA-Scenes, a dataset of 520 usage instances across 26 keywords for distinct scene identification; and (3) empirical evidence from two experiments suggesting that scenes are reliably identifiable across human observers (82.4% accuracy, +11.8 pp over text-only embeddings) and that our scene profiles more closely align with human interpretation of words in context than ATOMIC-based alternatives (86.4% preference across three semantic dimensions).

Structured road understanding of lane geometry, topology, and traffic element relationships is foundational to safe autonomous driving. While vision-language models (VLMs) offer promising semantic flexibility, they lack the geometric and relational grounding required for precise road reasoning. Conversely, traditional modular systems, e.g., HD maps and topological road graphs, provide structural precision but remain semantically rigid. To bridge this gap, we introduce the Combined Road Substrate (CRS), a graph-grounded framework that makes geometric road structure and open-vocabulary semantics jointly executable in a single representation. CRS enables the automatic generation of compositionally complex and linguistically varied question-answer pairs via recursive graph queries, augmented with a "grounding for free" mechanism that ensures logical traceability to specific map elements, and procedurally extracted chain-of-thought supervision traces. We demonstrate that state-of-the-art VLMs - including large, closed-source models - struggle significantly with structured road reasoning, yet training a small 2- or 4-billion-parameter model with as few as 20 to 80 CRS-enriched scenes yields stable gains in compositional reasoning tasks of varying depth. Analysis of model behavior via verifiable reasoning traces reveals a systematic shift in failure modes: whereas baseline models fail at relational scene understanding, CRS-trained models reduce failures to attribute recognition, suggesting that the primary bottleneck in road understanding is not model scale, but the absence of structured supervision.

Comprehensive situational awareness is essential for autonomous vehicles operating in safety-critical environments, as it enables the identification and mitigation of potential risks. Although recent Multimodal Large Language Models (MLLMs) have shown promise on general vision-language tasks, our findings indicate that zero-shot MLLMs still underperform compared to domain-specific methods in fine-grained, spatially grounded risk assessment. To address this gap, we propose DriveSafe, a framework for risk-aware scene understanding that leverages structured natural language descriptions. Specifically, our method first generates spatially grounded captions enriched with multimodal context, including motion, spatial, and depth cues. These captions are then used for downstream risk assessment, explicitly identifying hazardous objects, their locations, and the unsafe behaviors they imply, followed by actionable safety suggestions. To further improve performance, we employ caption-risk pairings to fine-tune a lightweight adapter module, efficiently injecting domain-specific knowledge into the base LLM. By conditioning risk assessment on explicit language-based scene representations, DriveSafe achieves significant gains over both zero-shot MLLMs and prior domain-specific baselines. Exhaustive experiments on the DRAMA benchmark demonstrate state-of-the-art performance, while ablation studies validate the effectiveness of our key design choices. Project page: https://cvit.iiit.ac.in/ research/projects/cvit-projects/drivesafe

Autoregressive video diffusion models enable open-ended generation through local attention and KV caching. However, existing training-free long-video optimization methods mainly focus on stable extension under a single prompt, making them difficult to handle interactive scenarios involving prompt switching, old scene forgetting, and historical scene recall. We identify the core bottleneck as the functional entanglement of historical KV states: stable anchors and recent dynamics are handled by the same cache policy, leading to outdated background contamination, delayed response to new prompts, and loss of long-range memory. To address this issue, we propose Echo-Forcing, a training-free scene memory framework specifically designed for interactive long video generation with three core mechanisms: (1) Hierarchical Temporal Memory, which decouples stable anchors, compressed history, and recent windows under relative RoPE; (2) Scene Recall Frames, which compresses historical scenes into spatially structured KV representations to support long-term recall; and (3) Difference-aware Memory Decay, which adaptively forgets conflicting tokens according to the discrepancy between old and new scenes. Based on these designs, Echo-Forcing uniformly supports smooth transitions, hard cuts, and long-range scene recall under a bounded cache budget. Extensive evaluations on VBench-Long further demonstrate that Echo-Forcing achieves the best overall performance in both long-video generation and interactive video generation settings. Our code is released in https://github.com/mingqiangWu/Echo-Forcing

General scene perception has progressed from object recognition toward open-vocabulary grounding, part localization, and affordance prediction. Yet these capabilities are often realized as isolated predictions that localize objects, parts, or interaction points without capturing the structured dependencies needed for interaction-oriented scene understanding. To address this gap, we introduce Hierarchical Scene Parsing, an interaction-oriented parsing task that represents physical scenes as explicit scene -> object -> part -> affordance hierarchies with cross-level bindings. We instantiate this task with SceneParser, a VLM-based parser trained for unified hierarchical generation with structural-completion pseudo labels and curriculum learning. To support training and evaluation, we construct SceneParser-Bench, a large-scale benchmark built with a scalable hierarchical data engine, containing 110K training images, a 5K validation split, 777K objects, 1.14M parts, 1.74M affordance annotations, and 1.74M valid object-part-affordance chain instances. We further introduce Level-1 to Level-3 conditional metrics and ParseRate to evaluate localization, cross-level binding, and hierarchical completeness. Experiments show that existing MLLMs and perception-stitching pipelines struggle with hierarchical parsing on our SceneParser-Bench, while SceneParser achieves stronger structure-aware performance. Besides, ablations, evaluations on COCO and AGD20K, and a downstream planning probe demonstrate that our SceneParser is compatible with conventional tasks and provides an actionable representation for visual understanding.