cs.CV（2026-08-03）

📊 共 68 篇论文 | 🔗 18 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (24 🔗7) 支柱二：RL算法与架构 (RL & Architecture) (17 🔗3) 支柱三：空间感知与语义 (Perception & Semantics) (16 🔗4) 支柱七：动作重定向 (Motion Retargeting) (3 🔗1) 支柱一：机器人控制 (Robot Control) (3 🔗1) 支柱六：视频提取与匹配 (Video Extraction) (2) 支柱四：生成式动作 (Generative Motion) (1 🔗1) 支柱五：交互与反应 (Interaction & Reaction) (1 🔗1) 支柱八：物理动画 (Physics-based Animation) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (24 篇)

#题目一句话要点标签🔗⭐
1 Few-Shot Concept Prompt Learning for Segmentation Foundation Models via Visual Grounding 提出少样本概念提示学习以解决医学图像分割问题 foundation model visual grounding
2 Illuminating Visual Identity in Universal Multimodal Embeddings 提出统一视觉身份辨别方法以提升多模态嵌入能力 large language model multimodal ✅
3 Generative AI and Foundation Models in Medical Image 探讨生成式AI与基础模型在医学影像中的应用 large language model foundation model
4 UEmbed: Unified Sparse and Dense Multimodal Embeddings 提出UEmbed以解决多模态稀疏与密集嵌入统一问题 multimodal
5 Implicit Neural Representations for Multimodal Longitudinal Image Imputation and Interpolation 提出条件隐式神经表示以解决多模态MRI图像插补问题 multimodal
6 A General-Purpose VLM Can Teach an Astronomy Foundation Model to Better Recognize Galaxy Morphology 提出通用VLM以提升天文学基础模型的星系形态识别能力 foundation model ✅
7 Deep Multimodal Fusion Detection through Spatial Mask and Channel Fusion 提出基于注意力驱动的互补重采样框架以提升跨模态目标检测 multimodal
8 MIEScore: Human-Aligned Evaluation for Multi-Source Image Editing 提出MIEScore以解决多源图像编辑评估问题 large language model multimodal instruction following ✅
9 SPECTRA: Band-Routed Embedding and Stage-Wise LoRA for Cross-Sensor Fine-Tuning of Geospatial Foundation Models 提出SPECTRA以解决地理基础模型的光谱不匹配和适应成本问题 foundation model ✅
10 FAU at ImageCLEF 2026 Task on Multimodal Reasoning Robust Candidate Scoring and Concise Multilingual Visual Answering 提出多模态推理系统以解决视觉问答任务 multimodal
11 Mitigating Visual Degradation in MLLMs via Spatial-Spectral Visual Anchor Learning 提出空间-频谱视觉锚学习以解决MLLM视觉感知退化问题 large language model foundation model multimodal ✅
12 MoRAL: Sensor-Grounded BEV Reasoning for Compact VLMs toward Edge-Oriented Autonomous Driving 提出MoRAL以解决边缘计算平台上VLM的空间推理问题 multimodal chain-of-thought
13 ET-Prune: Evidence-Aware Dynamic Budgeting for Visual Token Pruning in Text-Rich MLLMs 提出ET-Prune以解决文本丰富输入的视觉令牌修剪问题 large language model multimodal
14 SVGEval: A Vision-Grounded Framework for Perceptual-Quality Benchmarking and Evaluation in Text-to-SVG Generation 提出SVGEval以解决SVG生成评估的可靠性问题 multimodal visual grounding
15 Decoupling semantics from vision: A framework for faithful visual-text compression evaluation 提出新评估框架以解决视觉文本压缩评估问题 large language model multimodal
16 GEOID-Flood: A Large-Scale Multi-Modal Benchmark Dataset for Flood Segmentation 提出GEOID-Flood以解决洪水分割数据集不足问题 foundation model ✅
17 Messages, Not Tokens: Grounded Coresets for Faithful VLM Compression 提出基于消息的压缩方法以解决视觉语言模型的效率问题 multimodal
18 Invisible Ink Threats: Adversarial Goals Behind Legitimate Tasks in Computer-Use Agents 提出II-Bench以应对计算机使用代理中的隐形墨水威胁 large language model
19 LongHorizon-Harness: Advancing Long-Horizon Agents for Real-World Tasks 提出LongHorizon-Harness以解决长时间任务状态管理问题 large language model
20 CultureVidBench: Benchmarking Cultural Understanding in Text-to-Video Generation 提出CultureVidBench以解决文本到视频生成中的文化理解问题 multimodal
21 Parameter-Dynamic Adaptive Fusion and Calibration Network for RGBT Tracking 提出PAFCNet以解决RGBT跟踪中的参数固定问题 multimodal
22 IDraw: Artist Verification from Digital Drawing Images 提出IDraw以解决数字绘画作者验证问题 multimodal
23 Learning Where to Look and How to Judge: Resolution-agnostic Image Quality Assessment with Quality-aware Saliency 提出ReLIQS以解决无参考图像质量评估中的分辨率依赖问题 multimodal
24 UniSim-SLAM: Feed-Forward SLAM with Unified Sim(3) Optimization 提出UniSim-SLAM以解决SLAM中的几何不一致性问题 foundation model ✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (17 篇)

#题目一句话要点标签🔗⭐
25 DeepVoyager-VL: Incentivizing Vision-in-the-Loop Search for Long-Horizon Multimodal Agents 提出DeepVoyager-VL以解决多模态长时间搜索问题 reinforcement learning large language model multimodal
26 DF$^3$: World Modeling via Decoder-Free Feature Forecasting in Autonomous Navigation 提出DF$^3$以解决自主导航中的世界建模问题 world model world models foundation model
27 AdaThinkV: Adaptive Thinking for Token-Efficient Video Reasoning 提出AdaThinkV以解决视频推理中的令牌效率问题 reinforcement learning large language model multimodal ✅
28 GeoCore-9B: Towards Geo-Aware Generative Foundation Models in Earth Observation 提出GeoCore-9B以解决地球观测生成模型的局限性 flow matching foundation model
29 WorldExam: Benchmarking World Models from Apparent Appearance to Inherent Reactivity 提出WorldExam基准以评估世界模型的内在反应能力 world model world models
30 ReMiX-MAE: Learning Missing-Channel Cross-Modal Representations from RGB-Only Clinical Facial Videos for Sympathetic-Mediated Pain Assessment 提出ReMiX-MAE以解决临床面部视频缺失通道问题 masked autoencoder MAE multimodal
31 ASTRA: Asynchronous Spatio-Temporal Reconstruction via Trajectory Alignment 提出ASTRA以解决动态场景重建中的时间异步问题 MAE gaussian splatting splatting
32 StyleForge: Indoor Furniture Styling by Counterfactual Reasoning in a Hypergraph Field 提出StyleForge以解决室内家具风格一致性问题 preference learning large language model multimodal
33 SPARE: Structural Parameter-Free Affinity Regularization for Flow Matching 提出SPARE以解决去噪扩散变换器训练收敛慢的问题 flow matching classifier-free guidance
34 VR3D: View-Robust 3D Representation Learning for Aerial-Ground Person Re-Identification 提出VR3D框架以解决空中与地面行人重识别中的视角变化问题 representation learning 3D reconstruction
35 PhyCheck: Fine-Grained Evidence-Grounded Dataset for Physical Law Understanding in Video-LLMs 提出PhyCheck以解决视频语言模型对物理规律理解不足的问题 world model world models large language model
36 Loop-Mamba: A Loop Mamba with Degradation-Aware and Shared Memory for Old Photo Restoration 提出Loop-Mamba以解决老照片修复中的多重退化问题 Mamba
37 VC-Tooler: Learning Compositional and Adaptive Visual Tool Use 提出VC-Tooler以解决视觉工具使用的适应性与组合性问题 reinforcement learning multimodal ✅
38 SpatialQuery: Benchmarking Geometry-Grounded Multi-Instance Spatial Reasoning in Vision-Language Models 提出SPATIALQUERY以解决视觉语言模型的空间推理问题 MAE chain-of-thought ✅
39 Linear Multi-Timescale Retention as a Memory-Efficient Vision-Language Bridge 提出线性多时间尺度保留模块以解决视觉语言模型的内存瓶颈问题 linear attention scene understanding
40 Same Semantics, Different Paths: Self-Improving Alignment for Vision-Text Compression 提出SPIRAL框架以解决视觉文本压缩中的路径不一致问题 DPO distillation
41 PartMat: Material-Aware 3D Part Decomposition with a Single Global Latent 提出PartMat以解决3D物体材料边界分解问题 reinforcement learning flow matching

🔬 支柱三：空间感知与语义 (Perception & Semantics) (16 篇)

#题目一句话要点标签🔗⭐
42 UniqueSplat: View-conditioned 3D Gaussian Splatting for Generalizable 3D Reconstruction 提出UniqueSplat以解决视角适应性不足的3D重建问题 3D gaussian splatting 3D reconstruction gaussian splatting
43 DerainSplat: Feed-Forward Clean 3D Gaussian Splatting from Sparse Rainy Views 提出DerainSplat以解决稀疏雨景下3D场景重建问题 3D gaussian splatting 3DGS gaussian splatting
44 DecoupleGS: Interactive 3D Gaussian Splatting for End-to-End Autonomous Driving Testing 提出DecoupleGS以解决E2E自动驾驶测试中的动态场景问题 3D gaussian splatting 3DGS gaussian splatting
45 StreamSplat: Streaming Feed-Forward 3D Gaussian Splatting 提出StreamSplat以解决在线3D场景更新问题 3D gaussian splatting 3DGS gaussian splatting
46 D^2-4DGS: Dual-Depth Guided Sparse-Camera 4D Gaussian Splatting 提出D^2-4DGS以解决稀疏相机动态4D高斯渲染问题 monocular depth metric depth gaussian splatting
47 GIFT: Geometry-Invariant Fine-Tuning for Non-Lambertian Monocular Depth Estimation 提出GIFT以解决非朗伯面单目深度估计问题 depth estimation monocular depth foundation model
48 FAST-GS: Frequency Aware Space-time Gaussian Splatting for Photorealistic Dynamic Novel View Synthesis 提出频率感知时空高斯点云以解决动态场景合成问题 3D reconstruction gaussian splatting splatting
49 EOVSAM: Efficient Open-Vocabulary Segmentation with SAM 3 in One Pass 提出EOVSAM以解决SAM 3计算开销大的问题 open-vocabulary open vocabulary ✅
50 PhotoHOI: Synthesizing 3D Hand-Object Interactions from a Single RGB Photograph 提出PhotoHOI以解决从单张RGB照片合成3D手-物体交互的问题 open-vocabulary open vocabulary affordance
51 G-Skin: Learning to Bind 3D Gaussians with Generative Visual Priors 提出G-Skin以解决3D高斯模型绑定与动画生成问题 3D gaussian splatting gaussian splatting splatting ✅
52 InfiniSplat: Implicit Gaussian Decoding for Large-Baseline Monocular View Synthesis 提出InfiniSplat以解决单图像3D高斯渲染问题 3D gaussian splatting 3DGS gaussian splatting ✅
53 GSRAIN: Physically Calibrated High-/Low-Frequency Rainfall Synthesis for 3D Gaussian Driving Scenes 提出GSRAIN以解决自主驾驶场景中降雨模拟的物理可控性问题 3D gaussian splatting 3DGS gaussian splatting
54 Action-grounded tissue affordance enables anticipatory auto-framing that lowers surgeon cognitive workload during laparoscopic surgery 提出DiffeoAfford框架以降低腹腔镜手术中外科医生的认知负担 affordance
55 CLEAR: Conflict-aware Learning via Evidence-guided Adaptive Routing for Unified Sparse-View 3D Gaussian Super-Resolution 提出CLEAR以解决稀疏视图3D高斯超分辨率问题 3D gaussian splatting gaussian splatting splatting
56 CHOW-SLAM: Compact Hybrid Representation with Complementary Overlap Window Optimization for RGB-D SLAM 提出CHOW-SLAM以解决RGB-D SLAM中的空间与时间约束问题 NeRF neural radiance field scene reconstruction ✅
57 MoCRA: Mixture of Compositional Rank-1 Atoms for 4K All-in-One Video Restoration 提出MoCRA以解决4K视频恢复中的多重退化问题 optical flow

🔬 支柱七：动作重定向 (Motion Retargeting) (3 篇)

#题目一句话要点标签🔗⭐
58 Sen-Cap: Sensor-Flexible and Noise-Resilient Human Motion Capture via LiDAR-Camera Integration 提出Sen-Cap以解决多模态传感器对齐及噪声鲁棒性问题 human motion
59 Motion Beyond Morphology: Bootstrapping Cross-Category Motion Transfer from Abstract Motion Representations 提出超越形态的运动转移方法以解决视频动画问题 motion representation ✅
60 Extended Field of View Analysis for VideoGAN-based Trajectory Generation 提出基于VideoGAN的轨迹生成方法以解决交通行为复杂性问题 spatial relationship

🔬 支柱一：机器人控制 (Robot Control) (3 篇)

#题目一句话要点标签🔗⭐
61 Two Sides of the Same Coin: Co-Evolving Search for Cross-Task Attacks on Vision-Language Models 提出进化计算引导的跨任务攻击框架以增强视觉语言模型的鲁棒性 trajectory optimization multimodal
62 Open-Set Visual Text Forensics via Sparse-Constraint Rectified Flow 提出生成检测器以解决开放集视觉文本取证问题 manipulation flow matching
63 SpatioLM: Towards General Physical Spatial Intelligence in Vision-Language Models 提出SpatioLM以解决视觉语言模型的空间推理问题 manipulation ✅

🔬 支柱六：视频提取与匹配 (Video Extraction) (2 篇)

#题目一句话要点标签🔗⭐
64 Dynamic Resolution Routing for Efficient Egocentric Grounding 提出SmartRes框架以解决自我中心视觉定位中的高分辨率输入问题 egocentric Ego4D large language model
65 HiResNets: Native Full-HD Video Recognition with Foveal Residual Streams 提出HiResNets以解决高分辨率视频识别中的内存瓶颈问题 egocentric

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
66 GenPrior: Unleashing Text-to-Motion Generative Priors for Zero-Shot Skeleton-based Action Recognition 提出GenPrior以解决零样本骨架动作识别中的语义-运动差距问题 text-to-motion ✅

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

#题目一句话要点标签🔗⭐
67 UniMoCa: Unifying Motion and Camera Controls as Visual Proxies for Faithful Human Video Generation 提出UniMoCa以解决多人物视频生成中的运动与相机控制问题 multi-person interaction human motion ✅

🔬 支柱八：物理动画 (Physics-based Animation) (1 篇)

#题目一句话要点标签🔗⭐
68 Global-Scale Self-Supervised Spatiotemporal Learning for NDVI Time-Series Reconstruction 提出GloSSR框架以解决NDVI时间序列重建问题 spatiotemporal

⬅️ 返回 cs.CV 首页 · 🏠 返回主页