cs.CV（2026-06-25）

📊 共 46 篇论文 | 🔗 13 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (18 🔗6) 支柱二：RL算法与架构 (RL & Architecture) (14 🔗4) 支柱三：空间感知与语义 (Perception & Semantics) (7 🔗1) 支柱一：机器人控制 (Robot Control) (5) 支柱四：生成式动作 (Generative Motion) (2 🔗2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (18 篇)

#题目一句话要点标签🔗⭐
1 Position Rebinding Cache Reuse: Replay-Free Visual Revisiting for Interleaved Multimodal Reasoning 提出位置重绑定缓存重用以解决视觉证据重访问题 multimodal visual grounding
2 Ask, Solve, Generate: Self-Evolving Unified Multimodal Understanding and Generation via Self-Consistency Rewards 提出自演化框架以提升多模态理解与生成能力 multimodal
3 Paying More Attention to Visual Tokens in Self-Evolving Large Multimodal Models 提出VISE框架以解决视觉内容关注不足问题 multimodal ✅
4 HarmVideoBench: Benchmarking Harmful Video Understanding in Large Multimodal Models 提出HarmVideoBench以解决有害视频理解的评估问题 multimodal
5 On-board Remote-Sensing Foundation Models for Unsupervised Change Detection of Disaster Events 提出基于RSFM的无监督灾害事件变化检测方法 foundation model
6 Unison: Benchmarking Unified Multimodal Models via Synergistic Understanding and Generation 提出Unison基准以评估统一多模态模型的理解与生成能力 multimodal ✅
7 CORTEX: A Structured Reasoning Benchmark for Trustworthy 3D Chest CT MLLMs 提出CORTEX以解决3D胸部CT诊断推理不足问题 large language model multimodal
8 ReasonCLIP-58M: Visually Grounded Commonsense Reasoning Supervision for CLIP 提出ReasonCLIP-58M以解决CLIP模型在常识推理中的不足 large language model multimodal ✅
9 DiCoBench: Benchmarking Multi-Image Fine-Grained Perception via Differential and Commonality Visual Cues 提出DiCoBench以解决多图像细粒度感知评估问题 large language model multimodal
10 Perception, Verdict, and Evolution: Hindsight-Driven Self-Refining Forensics Agent for AI-Generated Image Detection 提出ForeAgent以解决深度伪造图像检测中的灵活性不足问题 large language model multimodal
11 See & Sniff: Learning Visuo-Olfactory Representations 提出See & Sniff框架以解决嗅觉与视觉的跨模态学习问题 multimodal
12 Pseudo-Text-Conditioned 3D Grounding DINO for Organ Localization in Abdominal CT 提出CT-3GDINO以解决腹部CT器官定位问题 multimodal
13 TraMP-LLaMA: Generative Interpretability with Decoupled Instruction Tuning for Facial Expression Quality Assessment 提出TraMP-LLaMA以解决面部表情质量评估的可解释性问题 multimodal ✅
14 Depth-Semantic Alignment and Affinity-Guided Fusion for Structured Radar Point Cloud Generation 提出基于视觉-雷达融合的点云生成方法以解决雷达点云稀疏问题 multimodal
15 Fine-tuning a multimodal large language model for clinician-grade autism behavioral scoring from short home videos 提出多模态大语言模型以解决自闭症行为评分问题 large language model multimodal
16 DMV-Bench: Diagnosing Long-Horizon Multimodal Agents' Visual Memory with Incidental Cue Injection 提出DMV-Bench以解决多模态智能体视觉记忆评估问题 multimodal ✅
17 Dismantling Pathological Shortcuts: A Causal Framework for Faithful LVLM Decoding 提出Fox框架以解决大规模视觉语言模型的幻觉问题 visual grounding ✅
18 TruEye: Fine-Grained Detection of AI-Generated Human Subjects in Images 提出TruEye以解决AI生成图像检测的挑战 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (14 篇)

#题目一句话要点标签🔗⭐
19 SatSplatDiff: Geometry-preserving generative refinement for high-fidelity satellite Gaussian Splatting 提出SatSplatDiff以解决卫星图像生成中的几何降解问题 MAE monocular depth 3D reconstruction ✅
20 Sculpting NeRF Geometry: Human-Preference Fine-Tuning of a 3D-Aware Face GAN 提出基于人类偏好的3D面部GAN微调方法以优化NeRF几何结构 reinforcement learning RLHF NeRF
21 Modeling Local, Global, and Cross-Modal Context in Multimodal 3D MRI 提出MICViT以解决多模态3D脑MRI的特征融合问题 representation learning multimodal
22 Not All Actions Are Equal: Rethinking Conditioning for Dexterous World Model 提出DexAC-WM以解决高自由度动作建模问题 world model world models
23 DanceOPD: On-Policy Generative Field Distillation 提出DanceOPD以解决图像生成模型多能力组合问题 flow matching distillation classifier-free guidance
24 PhysiFormer: Learning to Simulate Mechanics in World Space 提出PhysiFormer以模拟物理可行的3D物体运动 world model world models physically plausible ✅
25 ViQ: Text-Aligned Visual Quantized Representations at Any Resolution 提出ViQ框架以解决多模态表示中的信息损失问题 representation learning multimodal
26 Liquid Fusion of Heterogeneous Representations Towards General Salient Object Detection 提出液态融合网络解决通用显著目标检测问题 Mamba SSM state space model ✅
27 LiveEdit: Towards Real-Time Diffusion-Based Streaming Video Editing 提出实时扩散基础的视频编辑框架以解决流媒体编辑问题 distillation foundation model
28 Beyond MoCap: Scaling Motion Tokenizers with Synthetic Human Motion for Generative Modeling 提出合成人体运动扩展运动表示空间以解决生成建模问题 representation learning text-to-motion motion synthesis
29 Large Language Model Teaches Visual Students: Cross-Modality Transfer of Fine-Grained Conceptual Knowledge 提出LaViD框架以实现语言模型对视觉模型的知识迁移 distillation large language model multimodal ✅
30 ReWorld: Learning Better Representations for World Action Models 提出ReWorld以提升世界行动模型的表示学习效果 world model world models world action model
31 Qwen-Image-2.0-RL Technical Report 提出Qwen-Image-2.0-RL以提升图像生成与编辑质量 reinforcement learning RLHF distillation
32 MemoBench: Benchmarking World Modeling in Dynamically Changing Environments 提出MemoBench以解决动态环境下记忆一致性评估问题 world model world models

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

#题目一句话要点标签🔗⭐
33 Capacity-Controlled Multi-View Stylization of 3D Gaussian Splatting 提出容量控制框架以解决3D高斯点云多视角风格化问题 3D gaussian splatting 3DGS gaussian splatting
34 PanoImager: Geometry-Guided Novel View Synthesis and Reconstruction from Sparse Panoramic Views 提出PanoImager以解决稀疏全景图的3D重建问题 3DGS 3D reconstruction
35 FracEvent: Event-Camera Simulation via Fractional-Relaxation Pixel Dynamics 提出FracEvent以解决事件相机模拟中的像素动态问题 optical flow
36 Forget, Anticipate and Adapt: Test Time Training for Long Videos 提出帧遗忘网络以解决长视频测试时训练的计算挑战 depth estimation
37 CoIn: Comprehensive 2D-3D Inpainting with Gaussian Splatting Guidance 提出CoIn框架以解决3D场景修复问题 3DGS gaussian splatting splatting
38 Structured-Li-GS: Structured 3D Gaussians Splatting with LiDAR Incorporation and Spatial Constraints 提出结构化3D高斯点云渲染方法以提升重建质量 visual SLAM 3DGS 3D reconstruction
39 SemCityLoc: Aerial 6DoF Localization Using Semantic 3D City Models 提出SemCityLoc以解决无人机精确定位问题 monocular depth 3D reconstruction scene reconstruction ✅

🔬 支柱一：机器人控制 (Robot Control) (5 篇)

#题目一句话要点标签🔗⭐
40 OctoSense: Self-Supervised Learning for Multimodal Robot Perception 提出OctoSense以解决多模态机器人感知问题 quadruped masked autoencoder optical flow
41 Identifying the Unknown: Prompt-Free Open Vocabulary Anomaly Recognition for Robot-Object Interaction 提出AnomNOVIC以解决机器人对象交互中的开放词汇异常识别问题 humanoid humanoid robot masked autoencoder
42 PhysEditWorld: A Large-Scale Dataset Toward Physics-Editable World Models 提出PhysEditWorld以解决物理动态可控性不足的问题 manipulation world model world models
43 DanceDuo: Bridging Human Movement and AI Choreography 提出DanceDuo以解决舞蹈生成与音乐同步问题 humanoid
44 Perceptual 3D Simulation With Physical World Modeling 提出P3Sim以解决真实场景下3D变换预测问题 manipulation world model world models

🔬 支柱四：生成式动作 (Generative Motion) (2 篇)

#题目一句话要点标签🔗⭐
45 Anatomy-Guided Residual Motion Diffusion for Controllable 4D Cardiac MRI Synthesis 提出一种解剖引导的残差运动扩散方法以实现可控的4D心脏MRI合成 motion diffusion ✅
46 Calibrated Harmonic Overlaid Implicit Neural Representations for Multi-Dimensional Data 提出校准谐波叠加隐式神经表示以解决多维数据优化不稳定问题 physically plausible ✅

⬅️ 返回 cs.CV 首页 · 🏠 返回主页