Searcharxiv⌕ Search

arXiv subjects

Xiangbo Shu

Publications and source records attributed to Xiangbo Shu.

At least 19 recordsLinked to original sources
arXiv2026

Beyond Retrieval: Progressive Latent Memory Evolution for Streaming Video Understanding

Streaming video understanding requires multimodal large language models (MLLMs) to process continuous visual inputs and respond to user queries under strict causality and bounded memory. Existing approaches typically compress historical observations into an external memory bank and retrieve query-relevant evidence as additional visual context. Though effective, this store-and-retrieve paradigm keeps historical evidence as external visual context, preventing it from being internalized into a compact, evolving latent memory that can continuously guide streaming reasoning. To bridge this gap, we introduce LatentStream, a progressive latent working memory framework that shifts streaming memory from store-and-retrieve to retrieve-and-internalize. Specifically, LatentStream comprises three coordinated components. First, Query-agnostic Hierarchical Streaming Memory organizes visual history into short-, mid-, and long-term levels under a fixed memory budget through Jenks-guided adaptive consolidation. Once a query arrives, Hierarchical Latent Memory Evolution equips groups of latent memory tokens with progressively expanding memory receptive fields, enabling them to iteratively retrieve historical evidence from their corresponding scopes and internalize it into a compact, fixed-length latent memory. Finally, Progressive Confidence-guided Latent Memory Optimization constructs a hierarchical progression reward from group-wise predictive entropy and jointly refines the latent memory tokens and retrieved evidence, encouraging increasingly confident streaming reasoning. Extensive experiments demonstrate that LatentStream achieves new state-of-the-art results on existing online and offline video benchmarks.

cs.CV↗
arXiv2026

MedFG-VQA: Low-Frequency Memory and Graph Attention for Lightweight Medical VQA

Medical Visual Question Answering (Med-VQA) holds significant promise for clinical decision support, yet faces challenges due to limited annotated data and the high computational demands of existing large vision-language models. We propose MedFG-VQA, a lightweight framework that leverages a memory bank to augment DCT-based low-frequency features and employs graph-enhanced cross-attention for effective visual-textual alignment. Specifically, our approach features two key components: Frequency-Memory Fusion (FMF), which enhances low-frequency features by retrieving from a learnable memory bank built on DCT decomposition, and Graph-Aware Cross-Attention (GACA), which aligns visual-textual features via cross-attention and refines them through graph-convolutional aggregation. To address data scarcity, we construct SynMed-VQA, a large-scale synthetic dataset comprising over 2 million question-answer pairs across 9 imaging modalities and 10 major organs, generated with GPT-4o. Extensive experiments on SynMed-VQA and three other standard biomedical VQA benchmarks demonstrate that MedFG-VQA achieves competitive or superior performance compared to much larger models while maintaining significantly lower computational costs, highlighting its efficiency and potential for clinical deployment.

cs.CV↗
arXiv2026

PL-NBA: A Possession-level Universal Basketball Video Dataset Supporting Multiple Visual Understanding Tasks

Visual understanding in sports has emerged as a hot topic in computer vision in recent years. Most existing basketball video datasets adopt single action or activity as sample, which can neither preserve the temporal continuity of game events nor support complex tasks such as action anticipation. To address this issue, this paper constructs the first possession-level basketball video dataset (PL-NBA), in which each sample is composed of a complete NBA offensive possession. Collected from 60 NBA games, PL-NBA contains 11,000 valid offensive possession clips and 31,567 annotated events with player names, captions, event types and timestamps. Each video clip includes multiple events and preserves the continuity of events, which is helpful for analysis of tactic. Experiment is conducted on multiple visual understanding tasks, including event recognition, video captioning, temporal action localization and action anticipation. Experimental results show that existing methods achieve limited performance on above four tasks, demonstrating that PL-NBA is a challenging benchmark for sports video understanding.

cs.CV↗
arXiv2026

Dual Latent Memory in Vision-Language-Action Models for Robotic Manipulation

Mainstream Vision-Language-Action (VLA) models predict actions primarily from the current observation under a Markovian assumption, thus struggling with long-horizon, temporally dependent tasks. Existing memory-augmented VLAs either expand the observation window or retrieve history from the memory bank as auxiliary policy-side context. However, they leave memory outside the native latent embedding space of VLA reasoning, preventing historical experience from being fluidly interleaved with multimodal reasoning and action formation. To this end, we introduce LaMem-VLA, a latent-memory-native framework that reconstructs historical experience into latent memory tokens and directly interweaves them with VLA reasoning. At its core, LaMem-VLA introduces four coordinated components: (i) a curator that organizes historical experience into two complementary short-term and long-term memory vaults; (ii) a seeker that queries both vaults using the multimodal cognition to retrieve context-relevant evidence; (iii) a condenser that reconstructs the retrieved evidence into compact short-term and long-term latent memory tokens; and (iv) a weaver that injects these memory tokens with the current observation and instruction into one continuous embedding sequence. By representing, retrieving, and consuming historical experience entirely in the same continuous latent space, LaMem-VLA enables memory to directly participate in VLA reasoning and guide action generation under a bounded context. Extensive experiments on SimplerEnv and LIBERO demonstrate the superiority of our LaMem-VLA.

cs.RO↗
arXiv2026

Learning 3D Representations for Spatial Intelligence from Unposed Multi-View Images

Robust 3D representation learning forms the perceptual foundation of spatial intelligence, enabling downstream tasks in scene understanding and embodied AI. However, learning such representations directly from unposed multi-view images remains challenging. Recent self-supervised methods attempt to unify geometry, appearance, and semantics in a feed-forward manner, but they often suffer from weak geometry induction, limited appearance detail, and inconsistencies between geometry and semantics. We introduce UniSplat, a feed-forward framework designed to address these limitations through three complementary components. First, we propose a dual-masking strategy that strengthens geometry induction in the encoder. By masking both encoder and decoder tokens, and targeting decoder masks toward geometry-rich regions, the model is forced to infer structural information from incomplete visual cues, yielding geometry-aware representations even under unposed inputs. Second, we develop a coarse-to-fine Gaussian splatting strategy that reduces appearance-semantics inconsistencies by progressively refining the radiance field. Finally, to enforce geometric-semantic consistency, we introduce a pose-conditioned recalibration mechanism that interrelates the outputs of multiple heads by re-projecting predicted 3D point and semantic maps into the image plane using estimated camera parameters, and aligning them with corresponding RGB and semantic predictions to ensure cross-task consistency, thereby resolving geometry-semantic mismatches. Together, these components yield unified 3D representations that are robust to unposed, sparse-view inputs and generalize across diverse tasks, laying a perceptual foundation for spatial intelligence.

cs.CV↗
arXiv2026

Bilingual Text-to-Motion Generation: A New Benchmark and Baselines

Text-to-motion generation holds significant potential for cross-linguistic applications, yet it is hindered by the lack of bilingual datasets and the poor cross-lingual semantic understanding of existing language models. To address these gaps, we introduce BiHumanML3D, the first bilingual text-to-motion benchmark, constructed via LLM-assisted annotation and rigorous manual correction. Furthermore, we propose a simple yet effective baseline, Bilingual Motion Diffusion (BiMD), featuring Cross-Lingual Alignment (CLA). CLA explicitly aligns semantic representations across languages, creating a robust conditional space that enables high-quality motion generation from bilingual inputs, including zero-shot code-switching scenarios. Extensive experiments demonstrate that BiMD with CLA achieves an FID of 0.045 vs. 0.169 and R@3 of 82.8\% vs. 80.8\%, significantly outperforms monolingual diffusion models and translation baselines on BiHumanML3D, underscoring the critical necessity and reliability of our dataset and the effectiveness of our alignment strategy for cross-lingual motion synthesis. The dataset and code are released at \href{https://wengwanjiang.github.io/BilingualT2M-page}{https://wengwanjiang.github.io/BilingualT2M-page}

cs.CV↗
arXiv2026

Beyond Quadratic: Linear-Time Change Detection with RWKV

Existing paradigms for remote sensing change detection are caught in a trade-off: CNNs excel at efficiency but lack global context, while Transformers capture long-range dependencies at a prohibitive computational cost. This paper introduces ChangeRWKV, a new architecture that reconciles this conflict. By building upon the Receptance Weighted Key Value (RWKV) framework, our ChangeRWKV uniquely combines the parallelizable training of Transformers with the linear-time inference of RNNs. Our approach core features two key innovations: a hierarchical RWKV encoder that builds multi-resolution feature representation, and a novel Spatial-Temporal Fusion Module (STFM) engineered to resolve spatial misalignments across scales while distilling fine-grained temporal discrepancies. ChangeRWKV not only achieves state-of-the-art performance on the LEVIR-CD benchmark, with an 85.46% IoU and 92.16% F1 score, but does so while drastically reducing parameters and FLOPs compared to previous leading methods. This work demonstrates a new, efficient, and powerful paradigm for operational-scale change detection. Our code and model are publicly available.

cs.CV↗
arXiv2026

PCA-Seg: Revisiting Cost Aggregation for Open-Vocabulary Semantic and Part Segmentation

Recent advances in vision-language models (VLMs) have garnered substantial attention in open-vocabulary semantic and part segmentation (OSPS). However, existing methods extract image-text alignment cues from cost volumes through a serial structure of spatial and class aggregations, leading to knowledge interference between class-level semantics and spatial context. Therefore, this paper proposes a simple yet effective parallel cost aggregation (PCA-Seg) paradigm to alleviate the above challenge, enabling the model to capture richer vision-language alignment information from cost volumes. Specifically, we design an expert-driven perceptual learning (EPL) module that efficiently integrates semantic and contextual streams. It incorporates a multi-expert parser to extract complementary features from multiple perspectives. In addition, a coefficient mapper is designed to adaptively learn pixel-specific weights for each feature, enabling the integration of complementary knowledge into a unified and robust feature embedding. Furthermore, we propose a feature orthogonalization decoupling (FOD) strategy to mitigate redundancy between the semantic and contextual streams, which allows the EPL module to learn diverse knowledge from orthogonalized features. Extensive experiments on eight benchmarks show that each parallel block in PCA-Seg adds merely 0.35M parameters while achieving state-of-the-art OSPS performance.

cs.CV↗
arXiv2026

Spatio-temporal Decoupled Knowledge Compensator for Few-Shot Action Recognition

Few-Shot Action Recognition (FSAR) is a challenging task that requires recognizing novel action categories with a few labeled videos. Recent works typically apply semantically coarse category names as auxiliary contexts to guide the learning of discriminative visual features. However, such context provided by the action names is too limited to provide sufficient background knowledge for capturing novel spatial and temporal concepts in actions. In this paper, we propose DiST, an innovative Decomposition-incorporation framework for FSAR that makes use of decoupled Spatial and Temporal knowledge provided by large language models to learn expressive multi-granularity prototypes. In the decomposition stage, we decouple vanilla action names into diverse spatio-temporal attribute descriptions (action-related knowledge). Such commonsense knowledge complements semantic contexts from spatial and temporal perspectives. In the incorporation stage, we propose Spatial/Temporal Knowledge Compensators (SKC/TKC) to discover discriminative object-level and frame-level prototypes, respectively. In SKC, object-level prototypes adaptively aggregate important patch tokens under the guidance of spatial knowledge. Moreover, in TKC, frame-level prototypes utilize temporal attributes to assist in inter-frame temporal relation modeling. These learned prototypes thus provide transparency in capturing fine-grained spatial details and diverse temporal patterns. Experimental results show DiST achieves state-of-the-art results on five standard FSAR datasets.

cs.CV↗
arXiv2026

Taming SAM3 in the Wild: A Concept Bank for Open-Vocabulary Segmentation

The recent introduction of \texttt{SAM3} has revolutionized Open-Vocabulary Segmentation (OVS) through \textit{promptable concept segmentation}, which grounds pixel predictions in flexible concept prompts. However, this reliance on pre-defined concepts makes the model vulnerable: when visual distributions shift (\textit{data drift}) or conditional label distributions evolve (\textit{concept drift}) in the target domain, the alignment between visual evidence and prompts breaks down. In this work, we present \textsc{ConceptBank}, a parameter-free calibration framework to restore this alignment on the fly. Instead of adhering to static prompts, we construct a dataset-specific concept bank from the target statistics. Our approach (\textit{i}) anchors target-domain evidence via class-wise visual prototypes, (\textit{ii}) mines representative supports to suppress outliers under data drift, and (\textit{iii}) fuses candidate concepts to rectify concept drift. We demonstrate that \textsc{ConceptBank} effectively adapts \texttt{SAM3} to distribution drifts, including challenging natural-scene and remote-sensing scenarios, establishing a new baseline for robustness and efficiency in OVS. Code and model are available at https://github.com/pgsmall/ConceptBank.

cs.CV↗
arXiv2026

IGA-LWP: An Iterative Gradient-based Adversarial Attack for Link Weight Prediction

Link weight prediction extends classical link prediction by estimating the strength of interactions rather than merely their existence, and it underpins a wide range of applications such as traffic engineering, social recommendation, and scientific collaboration analysis. However, the robustness of link weight prediction against adversarial perturbations remains largely unexplored.In this paper, we formalize the link weight prediction attack problem as an optimization task that aims to maximize the prediction error on a set of target links by adversarially manipulating the weight values of a limited number of links. Based on this formulation, we propose an iterative gradient-based attack framework for link weight prediction, termed IGA-LWP. By employing a self-attention-enhanced graph autoencoder as a surrogate predictor, IGA-LWP leverages backpropagated gradients to iteratively identify and perturb a small subset of links. Extensive experiments on four real-world weighted networks demonstrate that IGA-LWP significantly degrades prediction accuracy on target links compared with baseline methods. Moreover, the adversarial networks generated by IGA-LWP exhibit strong transferability across several representative link weight prediction models. These findings expose a fundamental vulnerability in weighted network inference and highlight the need for developing robust link weight prediction methods.

cs.SI↗
arXiv2026

CLF-ULP: Cross-Layer Fusion-Based Link Prediction in Dynamic Multiplex UAV Networks

In complex Unmanned Aerial Vehicle (UAV) networks, UAVs can establish dynamic and heterogeneous links with one another for various purposes, such as communication coverage, collective sensing, and task collaboration. These interactions give rise to dynamic multiplex UAV networks, where each layer represents a distinct type of interaction among UAVs. Understanding how such links form and evolve is both of theoretical interest and of practical importance for the control and maintenance of networked UAV systems. In this paper, we first develop a dynamic multiplex network model for UAV networks to characterize their dynamic and heterogeneous link properties. We then propose a cross-layer fusion-based deep learning model, termed CLF-ULP, to predict future inter-UAV links based on historical topology data. CLF-ULP incorporates graph attention networks to extract topological features within each layer and perform a cross-layer attention fusion to capture inter-layer dependencies. Furthermore, a shared-parameter long short-term memory network is employed to model the temporal evolution of each layer. To improve embedding quality and link prediction performance, we develop a joint loss function that considers both intra-layer and inter-layer UAV adjacency. Extensive experiments on simulated UAV datasets under diverse mobility patterns demonstrate that CLF-ULP achieves state-of-the-art performance in predicting links within dynamic multiplex UAV networks.

cs.NI↗
arXiv2025

Spatiotemporal-Untrammelled Mixture of Experts for Multi-Person Motion Prediction

Comprehensively and flexibly capturing the complex spatio-temporal dependencies of human motion is critical for multi-person motion prediction. Existing methods grapple with two primary limitations: i) Inflexible spatiotemporal representation due to reliance on positional encodings for capturing spatiotemporal information. ii) High computational costs stemming from the quadratic time complexity of conventional attention mechanisms. To overcome these limitations, we propose the Spatiotemporal-Untrammelled Mixture of Experts (ST-MoE), which flexibly explores complex spatio-temporal dependencies in human motion and significantly reduces computational cost. To adaptively mine complex spatio-temporal patterns from human motion, our model incorporates four distinct types of spatiotemporal experts, each specializing in capturing different spatial or temporal dependencies. To reduce the potential computational overhead while integrating multiple experts, we introduce bidirectional spatiotemporal Mamba as experts, each sharing bidirectional temporal and spatial Mamba in distinct combinations to achieve model efficiency and parameter economy. Extensive experiments on four multi-person benchmark datasets demonstrate that our approach not only outperforms state-of-art in accuracy but also reduces model parameter by 41.38% and achieves a 3.6x speedup in training. The code is available at https://github.com/alanyz106/ST-MoE.

cs.CV↗
arXiv2025

OmniGaze: Reward-inspired Generalizable Gaze Estimation In The Wild

Current 3D gaze estimation methods struggle to generalize across diverse data domains, primarily due to i) the scarcity of annotated datasets, and ii) the insufficient diversity of labeled data. In this work, we present OmniGaze, a semi-supervised framework for 3D gaze estimation, which utilizes large-scale unlabeled data collected from diverse and unconstrained real-world environments to mitigate domain bias and generalize gaze estimation in the wild. First, we build a diverse collection of unlabeled facial images, varying in facial appearances, background environments, illumination conditions, head poses, and eye occlusions. In order to leverage unlabeled data spanning a broader distribution, OmniGaze adopts a standard pseudo-labeling strategy and devises a reward model to assess the reliability of pseudo labels. Beyond pseudo labels as 3D direction vectors, the reward model also incorporates visual embeddings extracted by an off-the-shelf visual encoder and semantic cues from gaze perspective generated by prompting a Multimodal Large Language Model to compute confidence scores. Then, these scores are utilized to select high-quality pseudo labels and weight them for loss computation. Extensive experiments demonstrate that OmniGaze achieves state-of-the-art performance on five datasets under both in-domain and cross-domain settings. Furthermore, we also evaluate the efficacy of OmniGaze as a scalable data engine for gaze estimation, which exhibits robust zero-shot generalization on four unseen datasets.

cs.CV↗
arXiv2025

ASTRA: Let Arbitrary Subjects Transform in Video Editing

While existing video editing methods excel with single subjects, they struggle in dense, multi-subject scenes, frequently suffering from attention dilution and mask boundary entanglement that cause attribute leakage and temporal instability. To address this, we propose ASTRA, a training-free framework for seamless, arbitrary-subject video editing. Without requiring model fine-tuning, ASTRA precisely manipulates multiple designated subjects while strictly preserving non-target regions. It achieves this via two core components: a prompt-guided multimodal alignment module that generates robust conditions to mitigate attention dilution, and a prior-based mask retargeting module that produces temporally coherent mask sequences to resolve boundary entanglement. Functioning as a versatile plug-and-play module, ASTRA seamlessly integrates with diverse mask-driven video generators. Extensive experiments on our newly constructed benchmark, MSVBench, demonstrate that ASTRA consistently outperforms state-of-the-art methods. Code, models, and data are available at https://github.com/XWH-A/ASTRA.

cs.CV↗
arXiv2025

DA-Occ: Direction-Aware 2D Convolution for Efficient and Geometry-Preserving 3D Occupancy Prediction in Autonomous Driving

Efficient and high-accuracy 3D occupancy prediction is vital for the performance of autonomous driving systems. However, existing methods struggle to balance precision and efficiency: high-accuracy approaches are often hindered by heavy computational overhead, leading to slow inference speeds, while others leverage pure bird's-eye-view (BEV) representations to gain speed at the cost of losing vertical spatial cues and compromising geometric integrity. To overcome these limitations, we build on the efficient Lift-Splat-Shoot (LSS) paradigm and propose a pure 2D framework, DA-Occ, for 3D occupancy prediction that preserves fine-grained geometry. Standard LSS-based methods lift 2D features into 3D space solely based on depth scores, making it difficult to fully capture vertical structure. To improve upon this, DA-Occ augments depth-based lifting with a complementary height-score projection that explicitly encodes vertical geometric information. We further employ direction-aware convolution to extract geometric features along both vertical and horizontal orientations, effectively balancing accuracy and computational efficiency. On the Occ3D-nuScenes, the proposed method achieves an mIoU of 39.3% and an inference speed of 27.7 FPS, effectively balancing accuracy and efficiency. In simulations on edge devices, the inference speed reaches 14.8 FPS, further demonstrating the method's applicability for real-time deployment in resource-constrained environments.

cs.CV↗
arXiv2025

MUST: Multi-Scale Structural-Temporal Link Prediction Model for UAV Ad Hoc Networks

Link prediction in unmanned aerial vehicle (UAV) ad hoc networks (UANETs) aims to predict the potential formation of future links between UAVs. In adversarial environments where the route information of UAVs is unavailable, predicting future links must rely solely on the observed historical topological information of UANETs. However, the highly dynamic and sparse nature of UANET topologies presents substantial challenges in effectively capturing meaningful structural and temporal patterns for accurate link prediction. Most existing link prediction methods focus on temporal dynamics at a single structural scale while neglecting the effects of sparsity, resulting in insufficient information capture and limited applicability to UANETs. In this paper, we propose a multi-scale structural-temporal link prediction model (MUST) for UANETs. Specifically, we first employ graph attention networks (GATs) to capture structural features at multiple levels, including the individual UAV level, the UAV community level, and the overall network level. Then, we use long short-term memory (LSTM) networks to learn the temporal dynamics of these multi-scale structural features. Additionally, we address the impact of sparsity by introducing a sophisticated loss function during model optimization. We validate the performance of MUST using several UANET datasets generated through simulations. Extensive experimental results demonstrate that MUST achieves state-of-the-art link prediction performance in highly dynamic and sparse UANETs.

cs.LG↗
arXiv2025

Hierarchical Relation-augmented Representation Generalization for Few-shot Action Recognition

Few-shot action recognition (FSAR) aims to recognize novel action categories with few exemplars. Existing methods typically learn frame-level representations for each video by designing inter-frame temporal modeling strategies or inter-video interaction at the coarse video-level granularity. However, they treat each episode task in isolation and neglect fine-grained temporal relation modeling between videos, thus failing to capture shared fine-grained temporal patterns across videos and reuse temporal knowledge from historical tasks. In light of this, we propose HR2G-shot, a Hierarchical Relation-augmented Representation Generalization framework for FSAR, which unifies three types of relation modeling (inter-frame, inter-video, and inter-task) to learn task-specific temporal patterns from a holistic view. Going beyond conducting inter-frame temporal interactions, we further devise two components to respectively explore inter-video and inter-task relationships: i) Inter-video Semantic Correlation (ISC) performs cross-video frame-level interactions in a fine-grained manner, thereby capturing task-specific query features and enhancing both intra-class consistency and inter-class separability; ii) Inter-task Knowledge Transfer (IKT) retrieves and aggregates relevant temporal knowledge from the bank, which stores diverse temporal patterns from historical episode tasks. Extensive experiments on five benchmarks show that HR2G-shot outperforms current top-leading FSAR methods.

cs.CV↗