cs.CV（2026-06-18）

📊 共 34 篇论文 | 🔗 9 篇有代码

🎯 兴趣领域导航

支柱二：RL算法与架构 (RL & Architecture) (13 🔗5) 支柱九：具身大模型 (Embodied Foundation Models) (10 🔗1) 支柱三：空间感知与语义 (Perception & Semantics) (7 🔗3) 支柱一：机器人控制 (Robot Control) (1) 支柱六：视频提取与匹配 (Video Extraction) (1) 支柱五：交互与反应 (Interaction & Reaction) (1) 支柱七：动作重定向 (Motion Retargeting) (1)

🔬 支柱二：RL算法与架构 (RL & Architecture) (13 篇)

#题目一句话要点标签🔗⭐
1 UNIEGO: Proxies as Mediators for Unified Egocentric Video Representation Learning 提出UNIEGO以解决单视角视频理解的局限性问题 representation learning distillation egocentric
2 QueryGaussian: Scalable and Training-Free Open-Vocabulary 3D Instance Retrieval 提出QueryGaussian以解决大规模场景中的3D实例检索问题 distillation open-vocabulary open vocabulary
3 SurgVista: Long-Horizon Surgical World Modeling with Plausible Instrument-Tissue Dynamics 提出SurgVista以解决手术世界建模中的动态一致性问题 policy learning world model world models
4 HilDA: Hierarchical Distillation with Diffusion for Advancing Self-Supervised LiDAR Pre-trainin 提出HilDA框架以解决自监督LiDAR预训练中的知识蒸馏问题 distillation scene flow spatiotemporal ✅
5 SpatialSV: Internalizing Interpretable 3D Spatial Awareness in MLLMs via Task-Oriented Visual Supervision 提出SpatialSV以解决多模态大语言模型的3D空间意识问题 representation learning distillation 3D reconstruction
6 Exploring Multi-Modal Large Language Models and Two-Stage Fine-Tuning for Fashion Image Retrieval 提出多模态大语言模型与两阶段微调以提升时尚图像检索 contrastive learning large language model
7 Current World Models Lack a Persistent State Core 提出WRBench以解决世界模型缺乏持久状态核心的问题 world model world models
8 Holo-World: Unified Camera, Object and Weather Control for Video World Model 提出Holo-World以解决视频世界模型中的控制孤立问题 world model world models ✅
9 CARE: Competence-Aware Reward Shaping for Adaptive Reasoning Length in Video-MLLMs 提出CARE框架以优化多模态视频推理中的自适应推理长度 reinforcement learning reward shaping multimodal ✅
10 SAM3 Self-Distillation for Fine-Grained GOOSE 2D Semantic Segmentation 提出SAM3自蒸馏方法以解决细粒度GOOSE 2D语义分割问题 distillation foundation model
11 Spatial-Aware Reduction Framework: Towards Efficient and Faithful Visual State Space Models 提出STORM框架以解决视觉状态空间模型中的效率与准确性问题 Mamba state space model
12 U$^2$Mamba: A Two-level Nested U-structure Mamba for Salient Object Detection 提出U$^2$Mamba以解决显著目标检测中的上下文信息不足问题 Mamba ✅
13 Distill Once, Adapt Life-Long: Exploring Dataset Distillation for Continual Test-Time Adaptation 提出DO-ALL框架以解决长期测试时间适应问题 distillation ✅

🔬 支柱九：具身大模型 (Embodied Foundation Models) (10 篇)

#题目一句话要点标签🔗⭐
14 ROSE: Benchmarking the Perception-to-Action Gap in Multimodal Models 提出ROSE基准以解决多模态模型的感知与行动差距问题 large language model multimodal
15 Confidence Calibration for Multimodal LLMs: An Empirical Study through Medical VQA 提出多策略融合审问方法以提升医疗多模态大语言模型的信心校准 large language model multimodal
16 3D-PLOT-LLM: Part-Level Object Tokens for 3D Large Language Models 提出3D-PLOT-LLM以解决3D对象部分识别问题 large language model multimodal
17 HEad and neCK TumOR (HECKTOR) 2025: Benchmark of Segmentation, Diagnosis, and Prognosis in Multimodal PET/CT 建立HECKTOR 2025基准以提升头颈肿瘤分析精度 multimodal
18 Multimodal Concept Bottleneck Models 提出多模态概念瓶颈模型以解决可解释性与泛化问题 multimodal
19 SARLO-80: Worldwide Slant SAR Language Optic Dataset 80cm 提出SARLO-80数据集以解决SAR与光学图像融合问题 foundation model multimodal ✅
20 Evaluating and Enhancing Negation Comprehension in Remote Sensing MLLMs 提出RS-Neg基准以提升遥感多模态大语言模型的否定理解能力 large language model multimodal
21 Timage: A Generative Text-in-Image Paradigm for Fine-Tuning Vision-Language Models 提出Timage以解决多模态模型空间推理问题 large language model multimodal
22 The Hidden Evolution of Disguised Visual Context inside the VLM 提出视觉令牌的隐藏演变以优化视觉语言模型的集成 large language model
23 NEST: Narrative Event Structures in Time for Long Video Understanding 提出NEST以解决长视频叙事结构理解问题 multimodal

🔬 支柱三：空间感知与语义 (Perception & Semantics) (7 篇)

#题目一句话要点标签🔗⭐
24 Geometry-Preserving in 3D Gaussian Splatting for LiDAR-Camera Extrinsic Calibration 提出几何保持方法以解决LiDAR-相机标定问题 3D gaussian splatting 3DGS gaussian splatting
25 ReA-OVCD: Reliability-Aware Open-Vocabulary Change Detection via Semantic and Spatial Refinement 提出ReA-OVCD以解决开放词汇变化检测中的可靠性问题 open-vocabulary open vocabulary ✅
26 See-and-Reach: Precise Vision-Language Navigation for UAVs within the Field of View 提出UAV-VLN-FOV以解决无人机精确视觉语言导航问题 3DGS egocentric VLN ✅
27 Occ-VLM: Occupancy Grounded Vision Language Model for Indoor Scene Understanding 提出Occ-VLM以解决室内场景理解中的3D几何感知问题 scene understanding large language model
28 VisDom: Sparse Novel View Synthesis with Visible Domain Constraint 提出VisDom以解决稀疏新视图合成中的几何恢复问题 gaussian splatting splatting NeRF
29 Triangular Consistency as a Universal Constraint for Learning Optical Flow 提出三角一致性约束以提升光流学习效果 optical flow
30 One-Shot Novel View and Pose Human Image Synthesis via 3D Prior Guided Diffusion Model 提出基于3D先验引导的扩散模型以解决单次人像合成问题 NeRF ✅

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
31 EventVLA: Event-Driven Visual Evidence Memory for Long-Horizon Vision-Language-Action Policies 提出EventVLA以解决长时间机器人操作中的记忆瓶颈问题 manipulation bi-manual vision-language-action

🔬 支柱六：视频提取与匹配 (Video Extraction) (1 篇)

#题目一句话要点标签🔗⭐
32 HumanScale: Egocentric Human Video Can Outperform Real-Robot Data for Embodied Pretraining 提出人类视频数据以超越真实机器人数据的预训练效果 egocentric large language model foundation model

🔬 支柱五：交互与反应 (Interaction & Reaction) (1 篇)

#题目一句话要点标签🔗⭐
33 Spectral Query-Key Product Weight Steering for Training-Free VLM Hallucination Mitigation 提出QK产品引导以解决视觉语言模型的幻觉问题 mutual attention multimodal

🔬 支柱七：动作重定向 (Motion Retargeting) (1 篇)

#题目一句话要点标签🔗⭐
34 CalTennis: Large Multi-View Tennis Video Dataset and Benchmark of Monocular-to-3D Pose Estimation 提出CalTennis数据集以解决单目到3D姿态估计的评估问题 human motion

⬅️ 返回 cs.CV 首页 · 🏠 返回主页