Searcharxiv⌕ Search

arXiv subjects

Kunyu Peng

Publications and source records attributed to Kunyu Peng.

At least 19 recordsLinked to original sources

SGR3 Model: Scene Graph Retrieval-Reasoning Model in 3D

3D scene graphs provide a structured representation of object entities and their relationships, enabling high-level interpretation and reasoning for robots while remaining intuitively understandable to humans. Existing approaches for 3D scene graph generation typically combine scene reconstruction with graph neural networks (GNNs). However, such pipelines require multi-modal data that may not always be available, and their reliance on heuristic graph construction can constrain the prediction of relationship triplets. In this work, we introduce a Scene Graph Retrieval-Reasoning Model in 3D (SGR3 Model), a training-free framework that leverages multi-modal large language models (MLLMs) with retrieval-augmented generation (RAG) for semantic scene graph generation. SGR3 Model bypasses the need for explicit 3D reconstruction. Instead, it enhances relational reasoning by incorporating semantically aligned scene graphs retrieved via a ColPali-style cross-modal framework. To improve retrieval robustness, we further introduce a weighted patch-level similarity selection mechanism that mitigates the negative impact of blurry or semantically uninformative regions. Experiments demonstrate that SGR3 Model achieves competitive performance compared to training-free baselines and on par with GNN-based expert models. Moreover, an ablation study on the retrieval module and knowledge base scale reveals that retrieved external information is explicitly integrated into the token generation process, rather than being implicitly internalized through abstraction.

cs.CV↗

Read, Parse, Describe: Unified Document Parsing with Visual Element Description Generation

Document understanding systems are typically evaluated as separate components via layout analysis, optical text recognition, reading-order prediction, and image captioning. This separation leaves scientific figures, captions, and visual semantics weakly connected, and fails to measure whether a model can produce a single page-level parse that is both structurally faithful and semantically accessible. We introduce Read-Parse-Describe, a unified dataset for document parsing with visual element description generation. Given a page image, models are required to output an ordered list of layout elements containing bounding boxes, semantic labels, reading-order indices, recognized textual content, and alternative text descriptions for visual elements. The benchmark is built from 2,402 documents, comprising 18,163 continuous pages, 125,218 layout elements, and 11,338 visual instances. Besides, we further propose a plug-and-play space-awareness enhancer that reuses multi-layer visual features and injects grid-aware spatial guidance into vision-language tokens for layout-aware generation. Experiments across end-to-end VLMs and pipeline parsers show that existing systems still struggle to jointly recover reading order and visual semantics, while our enhanced model achieves strong joint performance. These results highlight the need to move beyond isolated document parsing modules toward unified models that jointly read, parse, and describe complex documents. The new dataset and code will be made publicly available at https://github.com/yuheng98chen/RPD-Unidoc

cs.CV↗

Towards Multi-Source Domain Generalization for Sleep Staging with Noisy Labels

Automatic sleep staging is a multimodal learning problem involving heterogeneous physiological signals such as EEG and EOG, which often suffer from domain shifts across institutions, devices, and populations. In practice, these data are also affected by noisy annotations, yet label-noise-robust multi-source domain generalization remains underexplored. We present the first benchmark for Noisy Labels in Multi-Source Domain-Generalized Sleep Staging (NL-DGSS) and show that existing noisy-label learning methods degrade substantially when domain shifts and label noise coexist. To address this challenge, we propose FF-TRUST, a domain-invariant multimodal sleep staging framework with Joint Time-Frequency Early Learning Regularization (JTF-ELR). By jointly exploiting temporal and spectral consistency together with confidence-diversity regularization, FF-TRUST improves robustness under noisy supervision. Experiments on five public datasets demonstrate consistent state-of-the-art performance under diverse symmetric and asymmetric noise settings.

cs.LG↗

What Breaks Scribble-Supervised Segmentation? An Audit-Informed Benchmark for Scribble Annotation Robustness

Scribble-supervised segmentation is typically evaluated with a fixed annotation set, leaving robustness to distinct semantic and geometric imperfections poorly understood. Motivated by an audit of 12,031 manual PASCAL Visual Object Classes (VOC) 2012 scribbles and prior work, we introduce six controlled perturbation operators. Class corruption, spatial offset, and directional perturbation directly operationalize audit signals. Stroke-length reduction is motivated by prior scribble studies, while random dropout and fragmentation serve as diagnostic controls. Each operator reports requested and achieved changes. We evaluate manual and automatic scribbles across two image domains, two principal backbone settings, and multiple training pipelines, with an additional MiT-B1 PFA anchor evaluation. In the five-seed Tree Energy Loss (TEL)-ResNet101 VOC study, mean intersection over union (mIoU) drops from 70.08 to 1.46 under class corruption, compared with 35.57 for spatial offset, 52.20 for directional perturbation, and 63.24 for stroke-length reduction at their strongest settings. At the evaluated endpoints, semantic-anchor corruption produces the largest degradation, while geometric and support responses depend on the pipeline and dataset. Our benchmark provides a reproducible protocol for diagnosing annotation robustness beyond clean-set accuracy. The operator code is available at https://github.com/Wmhiram99/scribble-operator-benchmark.

cs.CV↗

LFX: Fusion-Form-Flexible Light Field Modeling for Dense Semantic Segmentation and Salient Object Detection

Two distinct light field (LF) observation forms, namely sub-aperture images (S.Aperture images) and focal stack (FS), correspond to two fusion settings: the central image with S.Aperture images (CS) and the central image with FS (CF). Existing LF perception methods are typically form-specific, while a modeling design that can be separately instantiated under either setting remains underexplored. To address this gap, we present LFX, a form-flexible LF perception framework. Here, ``X'' represents the LF observation form, indicating that the one modeling design can be separately instantiated under two fusion settings (CS and CF). Specifically, we propose Field-of-Parallax Angular Subspace Modeling (FoP-ASM), which maps the feature difference between each auxiliary observation and the central image to an observation-specific angular marker. This avoids pairwise modeling among auxiliary observations and form-specific geometric assumptions. Meanwhile, a shared manifold subspace and the associated regularization losses constrain the angular residuals within a common feature space, promoting semantic consistency across auxiliary observations. Extensive evaluations across three LF datasets show that LFX achieves state-of-the-art results across distinct LF forms. Under the S.Aperture images and FS settings, respectively, LFX reduces MAE by 12.12%/20.59% on DUTLF-V2 and by 6.15%/7.69% on HFUT, while achieving the best performance on UrbanLF-real. The source code will be made publicly available at https://github.com/FeiT-FeiTeng/LFX.

cs.CV↗

Seeing Together: Multi-Robot Cooperative Egocentric Spatial Reasoning with Multimodal Large Language Models

Multimodal Large Language Models (MLLMs) have made substantial progress in egocentric video understanding, but their ability to reason cooperatively from multiple embodied viewpoints remains largely unexplored. We study this problem through multi-robot cooperative dynamic spatial reasoning, where a model must answer spatial, temporal, visibility, and coordination questions by integrating synchronized egocentric videos from a team of moving robots. To support this setting, we introduce CoopSR, the first benchmark for this task, together with EgoTeam, a multi-robot egocentric QA dataset. EgoTeam contains 114,227 QA pairs spanning 19 question types, four difficulty tiers, and three team sizes in Habitat and iGibson, along with a real-world test set of around 2,326 QAs collected using two quadruped robots. We further propose SG-CoR (Spectral and Geometry-Aware Cooperative Reasoner), an MLLM framework for fine-grained cooperative spatial reasoning. SG-CoR combines dynamics-aware multi-robot frame sampling, spectral- and geometric-guided view fusion, and geometry- and motion-informed prompt distillation, enabling the model to benefit from privileged robot-pose supervision during training while requiring only egocentric videos at test time. Across 22 MLLM baselines, SG-CoR consistently improves cooperative reasoning, outperforming the strongest fine-tuned baseline by +3.87% on Habitat and +7.12% on iGibson. It also shows stronger generalization to unseen team sizes and real-world robot tests. Code can be found at https://github.com/KPeng9510/seeing-together.git.

cs.CV↗

InSight: A Mobile Assistant Helping Visually Impaired People Navigate Unfamiliar Environment

Navigating unfamiliar outdoor environments is one of the most demanding everyday tasks for people with visual impairments (PVI). However, existing navigation systems cannot simultaneously satisfy the needs of informative perception and precise localization during navigation. This paper explores the core navigation demands of PVI through a preliminary study. Based on the results of the pre-study, we present the design of InSight, a mobile assistant that combines scene descriptions with precise localization to support PVI navigating unfamiliar outdoor environments. The offline mode of InSight runs entirely on device for real-time navigation, while the online mode streams camera frames to a self-hosted server for arbitrary object detection and accurate localization. Finally, a user study with 21 visually impaired participants validates that InSight could assist users in navigating unfamiliar outdoor environments with increased success rate and confidence compared to mainstream navigation applications like Apple Maps.

cs.HC↗

VL-DocIR: A Benchmark for Vision-Based Long Document Information Retrieval

Vision-based document retrieval has improved rapidly, but current evaluation settings still emphasize short documents and single relevant pages. This leaves an important gap between leaderboard performance and realistic long-document retrieval, where the evidence needed to answer a query may span multiple pages or even multiple documents. We present VL-DocIR, a page-level benchmark for evidence-complete visual long-document retrieval. VL-DocIR is built from 29,641 HTML-born documents rendered into 388,548 page images from Wikipedia, arXiv, PubMed, and SEC proxy statements. It contains 271,760 queries across 23 domains and six evidence structures, covering single-page, within-document multi-page, and cross-document evidence configurations. Evidence annotations are grounded to rendered pages and HTML element identifiers, and queries are filtered with a cleaning pipeline targeting clarity, correctness, closedness, and the absence of explicit layout references. We evaluate leading single-vector and multi-vector visual retrievers together with diagnostic retrieval settings such as document-context scoring, query splitting, reranking, and round-robin ensembles. The strongest evaluated configuration reaches 69.3 nDCGAll@10 and 80.6 RecallAll@10 under structure-balanced Macro evaluation, but source- and evidence-structure breakdowns show that cross-document retrieval remains substantially harder than single-page retrieval, and that document-context scoring is especially helpful for within-document multi-page queries. VL-DocIR exposes these failure modes and provides a more demanding target for visual long-document retrieval. The new dataset and code will be made publicly available at https://yufanchen96.github.io/VL-DocIR.

cs.AI↗

$M^2$-Occ: Resilient 3D Semantic Occupancy Prediction for Autonomous Driving with Incomplete Camera Inputs

Semantic occupancy prediction enables dense 3D geometric and semantic understanding for autonomous driving. However, existing camera-based approaches implicitly assume complete surround-view observations, an assumption that rarely holds in real-world deployment due to occlusion, hardware malfunction, or communication failures. We study semantic occupancy prediction under incomplete multi-camera inputs and introduce $M^2$-Occ, a framework designed to preserve geometric structure and semantic coherence when views are missing. $M^2$-Occ addresses two complementary challenges. First, a Multi-view Masked Reconstruction (MMR) module leverages the spatial overlap among neighboring cameras to recover missing-view representations directly in the feature space. Second, a Feature Memory Module (FMM) introduces a learnable memory bank that stores class-level semantic prototypes. By retrieving and integrating these global priors, the FMM refines ambiguous voxel features, ensuring semantic consistency even when observational evidence is incomplete. We introduce a systematic missing-view evaluation protocol on the nuScenes-based SurroundOcc benchmark, encompassing both deterministic single-view failures and stochastic multi-view dropout scenarios. Under the safety-critical missing back-view setting, $M^2$-Occ improves the IoU by 4.36%. As the number of missing cameras increases, the robustness gap further widens; for instance, under the setting with five missing views, our method boosts the IoU by 6.67%. These gains are achieved without compromising full-view performance. The source code will be publicly released at https://github.com/qixi7up/M2-Occ.

cs.CV↗

RefAtomNet++: Advancing Referring Atomic Video Action Recognition using Multi-Trajectory Semantic Retrieval

Who is being described, where are they, and what are they doing? Referring Atomic Video Action Recognition (RAVAR) answers these questions jointly by grounding a natural-language reference to a person and recognizing that person's fine-grained atomic actions in complex multi-person videos. Progress in RAVAR is constrained by limited benchmark scale and weak alignment between fine-grained linguistic and scene cues and temporally coherent visual cues. We address both challenges with a new dataset and model. We introduce RefAVA++, a large-scale dataset comprising 2,950,920$ frames, 75,111 annotated person instances, and 80 atomic action categories. Its references describe appearance and spatial attributes while deliberately omitting action labels, requiring models to infer actions directly from visual cues. We further propose RefAtomNet++, which models complementary semantics at the holistic-sentence, partial-keyword, and scene-attribute levels. Fine-grained semantic cues retrieve aligned visual tokens across time to construct trajectories, which are aggregated through Mamba-based state-space modeling and fused using multi-hierarchical semantic-aligned cross-attention. This design enables accurate joint person localization and multi-label atomic action recognition. Equipped with the InternVideo2.5 backbone, RefAtomNet++ achieves 50.39%/51.14% mIoU, 59.21%/59.55% mAP, and 73.89%/76.59% AUROC on RefAVA, and 49.49%/50.33% mIoU, 62.33%/61.90% mAP, and 76.06%/76.10% AUROC on RefAVA++ validation/test sets, respectively. The dataset and code are available at https://github.com/KPeng9510/refAVA2.

cs.CV↗

FUSEye: Training-Light Fisheye Detection with Overlapping Views and Zero-Initialized Adapters

Fisheye cameras give mobile robots a single-sensor, low-cost view of their surroundings, yet the COCO-pretrained detectors that practitioners routinely reuse fail on them: strong radial distortion warps local image structure, while boundary compression shrinks objects to near-invisible sizes. Full fine-tuning closes much of the gap but requires abundant fisheye labels and compute. We present FUSEye, a training-light framework that turns a frozen-backbone COCO-pretrained extra-large YOLO26 detector (YOLO26-x) into a fisheye detector. FUSEye adds roughly 227k new parameters while updating the inserted modules and the pretrained detection head. It addresses the transfer gap at three causally linked levels. At the input level, overlapping grid view generation and box remapping (GridViews) enlarge compressed boundary regions. At the feature level, zero-initialized residual adapters (Z-Adapters) correct distortion-induced feature misalignment. At the decision level, learned cross-projection agreement fusion (AgreeFusion) promotes low-confidence detections only when they are supported by consistent evidence across multiple views. On the WoodScape surround-view fisheye benchmark, FUSEye raises YOLO26-x from 0.148 to 0.266 mAP50 and retains 84.3% fully fine-tuned accuracy. Moreover, randomly using only 25% of the labeled training images, FUSEye achieves 0.2597 mAP50, retaining 97.6% of its full-label performance. FUSEye also consistently improves YOLOv8-11 detectors, showing that the recipe is architecture-agnostic. Source code will be available at https://github.com/Su-wenya/FUSEye.

cs.CV↗

OmniAct3D: Leveraging Foundation Geometry and Evidence-Grounded Reasoning for Panoramic 3D Detection

Accurate 3D detection is essential for mobile embodied agents, while Vision Foundation Models (VFMs) offer transferable visual and geometric priors. Yet existing VFM-based 3D detectors rely on narrow-view monocular images or discrete perspective views, limiting coherent surround perception; equirectangular projection (ERP) instead encodes a continuous 360 scene in a single image. Direct transfer remains difficult because ERP organizes geometry and visual information differently, making object-relevant cues hard to model, localize, and preserve. We propose OmniAct3D, a framework that adapts perspective-trained VFM detectors to ERP while preserving transferable VFM priors. To resolve geometric mismatch, the ERP-Ray Geometry Adapter (ERGA-Ray) models spherical viewing rays and periodic spatial structure. To localize evidence in scene-wide context, the Visual-Action Reasoning Chain (VARC) grounds each hypothesis in relevant panoramic evidence and converts it into a structured geometric action. To recover local cues lost under fixed token budgets, the Appearance-Guided Heading Expert (AGHE) re-encodes object regions at higher resolution for heading estimation. Experiments show that OmniAct3D improves over the previous best 3D detector by 2.96 NDS points on Spheriverse and over the unadapted VFM baseline by 24.87 mAP points on PanoMMOcc. With target-specific geometry adaptation, VARC retains 95--98% of the same-configuration mAP, indicating reusable object-level 3D reasoning across sensing configurations. The source code will be made publicly available at https://github.com/FeiT-FeiTeng/OmniAct3D.

cs.CV↗

Does Local Video Understanding Transfer Across Encounters? The EgoGears Benchmark

Embodied systems must make knowledge acquired during one encounter usable in another despite changes in viewpoint, motion, and illumination. Yet aggregate cross-video accuracy conflates failures of local perception with failures to preserve observation identity, establish correspondence, and compose evidence, obscuring whether local video understanding actually transfers. We introduce EgoGears, a complementary single- and multi-video benchmark designed to diagnose this transition. It contains 567 single-video and 1,487 multi-video questions derived from 126 human-collected egocentric recordings covering 39 outdoor routes. Repeated traversals across movement speeds and lighting conditions ground comparisons in shared physical environments; 531 questions require alignment across independent recordings. Single-video questions measure the local visual, spatial, and motion evidence available to a model, while multi-video questions test whether evidence remains bound to the correct observation and can be composed into consistent route relationships. We report 29 single-video and 31 multi-video MLLM configurations across six model families in the main leaderboard. Among the 20 configurations evaluated comparably on both splits, every model performs worse on multi-video questions, with a mean decrease of 22.5 percentage points, and the gap persists when answer format and scoring are held fixed. The gap is not explained simply by additional videos or recording boundaries. The central bottlenecks are observation--evidence binding and ordered route-state tracking. The code and benchmark are publicly available at https://github.com/lei-qi-233/EgoGears.

cs.CV↗

HEIR: Learning Human-Entity Interactions with Functional Roles

Understanding human-entity interactions requires recovering each person-action event's participants, roles, and shared identities. This structure can support embodied agents by clarifying who acts on which entities and how, informing anticipation and coordination in shared environments. Standard HOI metrics score individual links, leaving complete event composition undermeasured. We introduce HEIR (Human-Entity Interactions with Functional Roles), an image benchmark for complete grounded participant-role sets across object, interpersonal, and self-directed interactions. It contains 18,730 images, six roles, 105 actions, and 437 nouns, with shared entities, role changes, and repeated fillers; 51.6% of images contain multiple actors and 62.1% contain multiple actions. HEIR pairs relation AP with complete-set AP and structural evaluation. We also introduce CoRISP (Compositional Role-aware Interaction Set Prediction), which uses shared entity identities to combine role-conditioned evidence and predict normalized participant-role sets. Cardinality and role-multiplicity potentials couple assignments through event size and role composition, with exact per-event normalization. Across 16 baselines, relation and complete-event rankings diverge even after aligning action weights. CoRISP leads the evaluated systems on repeated-role events and shared-participant images in HEIR by 2.87 and 3.82 Set mAP points, respectively. On V-COCO, CoRISP achieves 73.72/76.23 role AP and 61.06/68.59 complete-set AP on two-slot actions under Scenarios 1/2. These results show the value of learning and evaluating event composition alongside individual relations. The code and dataset are publicly available at https://github.com/Kratos-Wen/HEIR.

cs.CV↗

Authorization Closure Graph: Minimal Repair for LLM Agents with Evolving User Instructions

Tool-using large language model (LLM) agents increasingly perform state-changing actions that require user authorization. Yet existing approaches do not provide a principled mechanism for selectively updating prior authorization when only part of an instruction changes. To this end, we propose an Authorization-Closure-Graph (ACG)-based framework that represents authorization and its dependencies as an evolving, versioned state. ACG selectively invalidates authority affected by a revision while preserving unaffected portions of the authorization state, and computes a minimal repair that identifies only the missing evidence or authority required for execution. This enables agents to adapt to revised instructions while avoiding stale authority and unnecessary authorization requests. We evaluate ACG across three advanced LLMs in two natural tasks, and ACG consistently improves action safety rate and task success rate. Code is available at https://github.com/weiliang822/ACG.

cs.AI↗

ϕ-RIE: From Photorealistic Reconstruction to Interactive Environments

3D Gaussian Splatting (3DGS) can reconstruct a captured scene photorealistically, but the resulting representation does not by itself support physical interaction. Robot simulation instead requires object-level change, \textit{i.e.}, objects must move independently, make contact, and reveal previously occluded surroundings. This gap arises because object appearance may remain entangled with the background, while hidden object geometry and occluded background content may be unobserved. To address this challenge, we present ϕ-RIE, a Gaussian-native pipeline that converts selected objects into movable simulator assets while preserving the remaining reconstruction. Our key observation is that asset construction and source removal should be coupled, \textit{i.e.}, one object identity should define the movable asset and the scene content to remove and complete. Accordingly, Scene Observation supplies shared evidence to Coupled Scene Construction, which creates registered assets and completed background Gaussians for simulator-driven rendering in an Interactive Environment. This coupling preserves unedited Gaussians while aligning visual and physical state. On 50 ScanNet++ scenes, evidence-based selection and registration retry increase matched F1 at 20\,mm from 0.336 to 0.383 at fixed retention. Further tests demonstrate asset executability, manipulation gains over a single-generator baseline, and the visual cost of conversion. Together, these results demonstrate that \name\ enables interactive scene conversion.

cs.RO↗

Do LiDAR Language Models Really Understand Spatio-temporal Relationships?

Recent 4D LiDAR language models aim to reason about objects and their evolving spatial relationships. Yet, in our evaluation, always selecting the same option nearly matches the multiple-choice accuracy of two B4DL-derived configurations. We introduce LiDAR-Hallu, a geometry-referenced benchmark and diagnostic protocol with 10,000 questions across 150 nuScenes scenes. It covers object existence, ego-relative position, distance ordering, relative motion, and temporal localization, with explicit rules for selecting objects, comparing times, and determining reference answers. Our protocol combines fixed-answer and candidate-content controls, cross-scene pairs with identical prompts but opposite reference answers, and relation-specific recall. Analysis of 100,000 recorded responses reveals failures hidden by aggregate accuracy. Candidate duration alone makes temporal answers predictable without observing LiDAR. On paired questions, the models frequently give the same answer to scenes requiring opposite answers. Relation-specific analysis further shows that both configurations miss every positive lateral-motion case across all tested conditions. Temporal-shuffle contrastive decoding provides little net improvement, as repairs are largely offset by new errors and the main failures persist. These results show that evaluating spatio-temporal reasoning requires testing whether models distinguish the queried physical relationships, rather than relying on individual-answer accuracy alone. The source code, checkpoints, and data are released at https://github.com/Awesome4D/4DMLLM_Hallucination_Bench.

cs.CV↗

Algebraic Consistency Alone Does Not Certify Temporal Structure in Latent Action Models

Latent action models infer a code for the transition between two frames of action-free video. Recent methods regularise this code to compose additively and reverse antisymmetrically, and report order-of-magnitude reductions in the resulting errors as a label-free certificate that the code has captured temporal structure. We show that this conclusion does not follow. Reconstruction drives the decoded transition toward a difference of state features, for which both identities hold for any pairing, a solution the metric cannot distinguish from one encoding nuisance state or a coordinate convention. Across five source domains, a trained but unconstrained counterpart already achieves 83-97% of the reduction relative to an untrained anchor. The residual fold is governed as much by the decoder family as by what is learned. A constrained model retrained after its temporal pairing is destroyed still reaches, in each domain, a lower error than the unconstrained model on real data. Downstream, preserving the temporal pairing yields no consistent advantage on LIBERO-GOAL or LIBERO-SPATIAL, and across the tested arms the code's mean linear action decodability falls as the algebraic error improves. We also test the most direct repair, a violation-contrastive objective that requires the algebra to fail on destroyed pairings: in the tested configurations it yields only a marginal separation within the reconstruction budget, on training and test triples alike. We recommend a validation protocol that these methods currently lack: a baseline-corrected metric, retraining on destroyed pairings, and a seed-budget analysis.

cs.CV↗