2821 papers
The paper proposes AtomicVLA, a unified planning-and-execution framework that utilizes a Skill-Guided Mixture-of-Experts architecture to dynamically compose atomic skill abstractions, thereby significantly improving scalability and performance in long-horizon robotic manipulation and continual learning tasks compared to existing monolithic VLA models.
GLASS is a novel unsupervised framework that establishes dense 3D shape correspondence across challenging non-isometric and inter-class scenarios by integrating geometric spectral analysis with semantic priors from vision-language foundation models, achieving state-of-the-art performance through view-consistent feature extraction, language-injected vertex descriptors, and a graph-assisted contrastive loss.
This paper proposes a Self-Critical Inference (SCI) framework that enhances the robustness of Large Vision-Language Models against language bias and sensitivity through multi-round counterfactual reasoning with textual and visual perturbations, alongside a new Dynamic Robustness Benchmark (DRBench) for model-specific evaluation.
This paper introduces Holi-Spatial, the first fully automated, large-scale, spatially-aware multimodal dataset constructed from raw video streams without human intervention, which provides 4 million high-quality 3D semantic annotations and spatial QA pairs to significantly enhance the training and performance of Vision-Language Models on spatial reasoning tasks.
Ref-DGS is an efficient, rasterization-based framework that achieves state-of-the-art novel view synthesis on reflective scenes by decoupling surface geometry from specular reflections using a dual Gaussian representation and a lightweight adaptive mixing shader, thereby avoiding the high computational cost of explicit ray tracing.
This paper introduces FusionRegister, a general and efficient cross-modality registration framework guided by visual priors that directly corrects misalignment within fused infrared and visible images, thereby enhancing detail alignment and robustness without requiring extensive pre-registration.
UniUncer is a lightweight, unified framework for end-to-end autonomous driving that jointly estimates and leverages uncertainty for both static map elements and dynamic agents through probabilistic regression, uncertainty-aware query fusion, and adaptive gating, thereby significantly improving trajectory accuracy and planning robustness with minimal computational overhead.
FrameVGGT addresses the unbounded memory growth in streaming Visual Geometry Transformers by introducing a frame-driven rolling explicit-memory framework that aggregates frame-level evidence into compact prototypes, enabling stable long-sequence 3D perception under strict memory budgets.
This paper introduces RoboPCA, a pose-centered affordance learning framework that jointly predicts task-appropriate contact regions and poses from human demonstrations via the Human2Afford data curation pipeline, enabling robots to effectively manipulate objects with improved consistency and generalization across tasks and categories.
This paper proposes CDA-VSR, a compressed-domain-aware online video super-resolution network that leverages motion vectors, residual maps, and frame types to achieve real-time, high-quality reconstruction with significantly reduced computational cost compared to state-of-the-art methods.
This paper introduces the Masked Motion Diffusion Model (MMDM), a diffusion-based framework equipped with a Kinematic Attention Aggregation mechanism that learns context-adaptive motion priors to effectively reconstruct, refine, and complete 3D human motion from incomplete or noisy data.
TDM-R1 introduces a novel reinforcement learning paradigm that enables few-step diffusion models to effectively incorporate non-differentiable rewards by decoupling surrogate reward learning from generator training, achieving state-of-the-art performance across various metrics and scaling to powerful models like Z-Image with only 4 inference steps.
The paper introduces PARSE, a framework utilizing part-level geometric relations encoded in Part-centric Assembly Graphs to resolve spatial ambiguities, which is validated through the creation of the PARSE-10K dataset and demonstrated to significantly enhance both object layout reasoning in vision-language models and the physical realism of generated 3D scenes.
To address the "spatial intelligence gap" where Vision-Language Models struggle with elementary 3D tasks despite strong logical reasoning, the paper introduces 3ViewSense, a framework that leverages an engineering-inspired "Simulate-and-Reason" mechanism to ground spatial understanding in orthographic views, significantly improving performance on occlusion-heavy counting and view-consistent reasoning benchmarks.
The paper proposes AR2-4FV, a novel framework for long-term language-guided referring in fixed-view videos that leverages a static background-derived Anchor Bank and a ReID-Gating mechanism to maintain identity continuity and accelerate re-capture during occlusions or scene exits, significantly outperforming existing baselines in re-capture rate and latency.
The paper proposes DECADE, an unsupervised diffusion model that achieves temporally consistent denoising of Rb-82 dynamic cardiac PET images without paired training data, effectively reducing noise while preserving quantitative accuracy for myocardial blood flow and flow reserve metrics.
This paper introduces MedQ-Deg, a comprehensive benchmark featuring 24,894 expert-calibrated question-answer pairs across 18 degradation types and 7 imaging modalities, which reveals that mainstream medical multimodal large language models suffer systematic performance drops and exhibit the "AI Dunning-Kruger Effect" of overconfidence under image quality degradations.
This paper proposes the Geometric Knowledge-Guided Federated Dual Knowledge Distillation (GK-FedDKD) framework to address data heterogeneity in remote sensing satellite imagery by leveraging aggregated geometric knowledge from local covariance matrices and a dual distillation process to significantly outperform state-of-the-art methods.
This paper proposes a novel style transfer method that represents images in the brushstroke domain rather than the traditional pixel domain, offering a more natural and visually superior approach to mimicking artistic styles.
The paper introduces OrdinalBench, a comprehensive benchmark dataset and evaluation framework designed to diagnose and expose the significant generalization limitations of Vision-Language Models in understanding ordinal numbers and performing sequential reasoning tasks involving large indices and complex paths.