cs.CV（2026-11-03）

📊 共 50 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) (16 🔗8) 支柱二：RL算法与架构 (RL & Architecture) (12) 支柱三：空间感知与语义 (Perception & Semantics) (11 🔗4) 支柱一：机器人控制 (Robot Control) (3 🔗1) 支柱四：生成式动作 (Generative Motion) (3) 支柱六：视频提取与匹配 (Video Extraction) (2) 支柱八：物理动画 (Physics-based Animation) (2) 支柱七：动作重定向 (Motion Retargeting) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (16 篇)

#题目一句话要点标签🔗⭐
1 What Does the Model See? Representation Validity in Multimodal LLM-Based Video Analysis 提出多模态输入有效性分析以优化视频分析模型 large language model multimodal
2 VIGOR: Visual Gain Ordering for Hallucination Mitigation in Multimodal Discrete Diffusion Language Models 提出VIGOR以解决多模态离散扩散语言模型的幻觉问题 multimodal ✅
3 PATTS: Parallel-Aware Test-Time Scaling for Thought-Image Generation in Multimodal Discrete Diffusion LLMs 提出PATTS以解决多模态离散扩散模型中的图像生成问题 multimodal ✅
4 VisSearch-Bench: Evaluating Long Horizon Visual Search in Multimodal Agents 提出VisSearch-Bench以评估多模态智能体的长时视觉搜索能力 multimodal
5 OCR in the Foundation Model Era: From Specialist Models to Unified Models 综述OCR在基础模型时代的发展与挑战 foundation model ✅
6 DenseTwins: Dense Modality Decoupling for Multimodal Self-Supervision 提出DenseTwins以解决多模态自监督学习中的信息对齐问题 multimodal ✅
7 Is Forgetting Unified? Exploring Concept Unlearning in Unified Multimodal Models 提出UniForget以解决统一多模态模型中的概念遗忘问题 multimodal
8 VidUEU-Agent: A Data-Curation Agent for Multimodal Understanding, Editing, and Unified Tasks 提出VidUEU-Agent以解决多模态学习数据稀缺问题 multimodal
9 Factorized MLLM Embeddings for Composed Image Retrieval 提出因子化MLLM嵌入以解决复合图像检索问题 large language model multimodal
10 NightIQA: A Comprehensive Dataset and Benchmark for MLLM-Based Nighttime Image Quality Assessment 提出NightIQA以解决夜间图像质量评估问题 large language model multimodal
11 Beyond Token Representations: Explicit Visual Object Grounding for Video Reasoning Segmentation 提出显式视觉对象定位框架以解决视频推理分割问题 large language model multimodal ✅
12 Anchored Rank-Adversarial Fine-Tuning Pretrained Visual Place Recognition Models 提出锚定排名对抗微调以提升视觉场所识别模型的鲁棒性 foundation model
13 Perturb, Repair, Verify: Self-Play Vision-Language Verifiers for Compositional Understanding 提出PROVE框架以解决视觉-语言验证中的静态监督问题 multimodal
14 Manifold Embedding of Deep Image Features for Image Matching via Neural Adjoint Maps 提出基于神经伴随映射的流形嵌入方法以解决图像匹配问题 foundation model ✅
15 A Universal Raw-IQ Radar Encoder for Cross-Sensor Human Sensing 提出通用原始IQ雷达编码器以解决跨传感器人类感知问题 foundation model ✅
16 Entropy-Gated Dynamic Rectification for Object Hallucination in Large Vision-Language Models 提出熵门控动态整流以解决大规模视觉-语言模型中的物体幻觉问题 multimodal ✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (12 篇)

#题目一句话要点标签🔗⭐
17 Behind Video Generation: Observations and Insights on Multimodal Understanding from Kling 提出Kling-MLLM以解决视频生成中的多模态理解问题 reinforcement learning foundation model multimodal
18 YANAMI: A Large-Scale Multimodal Dataset and Benchmark for Multiturn Stateful Interaction 提出YANAMI以解决多轮状态交互的多模态数据集不足问题 world model world models multimodal
19 MambaHand: Local Identity and Behavioral Sequence Learning with Mamba for Dynamic Hand Gesture Authentication 提出MambaHand以解决动态手势认证中的身份与行为序列学习问题 Mamba representation learning motion representation
20 Mamba3R: Fixed-Capacity Selective State Memory for Long-Sequence 3D Reconstruction 提出Mamba3R以解决长序列3D重建的内存限制问题 Mamba 3D reconstruction
21 StepX-OPD: Rethinking Multi-Teacher On-Policy Distillation in Diffusion Models 提出StepX-OPD以解决多教师策略蒸馏在扩散模型中的效率问题 distillation instruction following
22 DreamX: Video Diffusion Model is a Universal Latent Action Learner 提出DreamX以解决视频动作表示学习中的转移问题 policy learning cross-embodiment spatiotemporal
23 ReconDreamer#: Self-Evolving Closed-Loop Synergy of Reconstruction and Generation for Dynamic Driving Scenes 提出ReconDreamer#以解决动态驾驶场景重建与生成的协同问题 dreamer scene reconstruction
24 Multi-Path Mamba Diffusion for 3D Point Cloud Semantic Scene Completion 提出MPath Mamba扩散网络以解决3D点云语义场景补全问题 Mamba state space model
25 E2I-Flow: EEG-to-Image Reconstruction via Wavelet-Temporal Cross-Attention and Conditional Flow Matching 提出E2I-Flow以解决EEG信号到图像重建问题 flow matching representation learning
26 Polarimetric Scattering-Structure Contrastive Learning for Few-Shot HRRP Recognition 提出极化散射结构对比学习以解决少样本高分辨率HRRP识别问题 contrastive learning
27 Pixel-space video generation with U-JiT 提出U-JiT以解决像素空间视频生成的挑战 flow matching spatiotemporal
28 Cycle WAM: Coupling Action and Video through Cycle Consistency 提出Cycle WAM以解决视频与动作一致性问题 world action model world action models

🔬 支柱三：空间感知与语义 (Perception & Semantics) (11 篇)

#题目一句话要点标签🔗⭐
29 3DGS-VBench+ and 3DGSV-Assessor: A Multi-Dimensional Subjective Benchmark and a Reference-Flexible Objective Model for 3D Gaussian Splatting Video Quality Assessment 提出3DGS-VBench+和3DGSV-Assessor以解决3D高斯点视频质量评估问题 3D gaussian splatting 3DGS gaussian splatting ✅
30 IGSW: Multi-Weather Urban Scene Reconstruction with Intrinsic-Decoupled Gaussian Splatting 提出IGSW以解决多天气城市场景重建中的结构失真问题 gaussian splatting splatting scene reconstruction ✅
31 Phase-DGS: Phase-guided Dynamic Gaussian Splatting from Unsynchronized Multi-view Video 提出Phase-DGS以解决动态高斯点云重建中的时间同步问题 gaussian splatting splatting scene reconstruction
32 GSCV++: Lightweight 3D Gaussian Splatting Sequence Compression via Standard Video Codecs 提出GSCV++以实现轻量级3D高斯点云序列压缩 3D gaussian splatting gaussian splatting splatting
33 SPA-Q: Structure-Preserving Adaptive Post-Training Quantization for Monocular Depth Estimation 提出SPA-Q以解决单目深度估计中的量化挑战 depth estimation monocular depth Depth Anything
34 EndoS3:Endoscopic Semantic Splatting SLAM 提出EndoS3以解决内窥镜语义SLAM中的多任务挑战 gaussian splatting splatting scene understanding
35 iMED: A Multi-Endoscope Dataset for Surgical 3D Perception 提出iMED数据集以解决外科手术视觉基准中的视角限制问题 depth estimation monocular depth feature matching
36 BhopalMLS: A LiDAR Dataset for Semantic Scene Understanding in Heterogeneous Urban Environments 提出BhopalMLS数据集以解决城市环境语义理解问题 scene understanding foundation model ✅
37 Visibility-Aware Reference Generation for Benchmarking Visual 3D Reconstruction in Vegetation-Rich Environments 提出基于LiDAR的参考生成管道以评估植被丰富环境中的3D重建 depth estimation 3D reconstruction
38 DynOSeg: Dynamic Object Segmentation via Motion-Geometry Fusion 提出DynOSeg以解决动态物体分割问题 scene understanding optical flow ✅
39 SparseGraph-I2P: Consistency-Guided Hypothesis Selection for Efficient Image-to-Point-Cloud Registration 提出SparseGraph-I2P以解决图像与点云注册中的模糊配对问题 monocular depth

🔬 支柱一：机器人控制 (Robot Control) (3 篇)

#题目一句话要点标签🔗⭐
40 FloMo: 3D Scene Flow as a World Action Model Intermediate for Learning from Human Video 提出FloMo模型以提升机器人动作预测的样本效率 teleoperation world action model world action models
41 SPOD: Recovering Suppressed Temporal Evidence in Video-LLMs via Single-Pass Orthogonal Decomposition 提出SPOD以解决视频大语言模型中的时间推理问题 manipulation large language model ✅
42 Route Anything: A Study of Model and Policy Routing for General-Purpose AI Systems 提出Route Anything框架以优化通用人工智能系统中的模型和策略路由 manipulation

🔬 支柱四：生成式动作 (Generative Motion) (3 篇)

#题目一句话要点标签🔗⭐
43 Möbius: A Unified Motion-Language Model Bridging Understanding and Generation 提出Möbius以解决统一人类动作理解与生成的问题 text-to-motion motion generation motion latent
44 VISAGE: Vertex-based Intrinsic Surface Appearance GEneration via Dual-Stream Diffusion for Arbitrary 3D Heads 提出VISAGE以解决3D人头表面外观生成问题 physically plausible
45 CounterSketch: Physics-Grounded Counterfactual Reasoning for Sketch-to-Video Generation 提出CounterSketch框架以实现物理基础的草图到视频生成 physically plausible

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

#题目一句话要点标签🔗⭐
46 Sparse Anchors for Visual Multi-Session Registration in Changing Environments 提出稀疏锚点以解决变化环境中的视觉多会话注册问题 feature matching
47 DPAGait: Dynamic Pose-Knowledge-Aware Network for Robust Gait Recognition in the Wild 提出动态姿态知识感知网络以解决野外步态识别问题 SMPL

🔬 支柱八：物理动画 (Physics-based Animation) (2 篇)

#题目一句话要点标签🔗⭐
48 On the Learnability of Footwear Robustness: Footstep Recognition Under Unseen Conditions 提出BiStepNet以解决鞋类变异下的足迹识别问题 spatiotemporal
49 Using GeoAI and machine learning tools for consistent high-resolution land cover mapping based on time-series NAIP imagery 提出双轨工作流程以解决历史航空影像一致性问题 spatiotemporal

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

#题目一句话要点标签🔗⭐
50 Gradient-Guided Orientation Learning for Weakly Supervised Oriented Object Detection 提出梯度引导的方向学习以解决弱监督定向目标检测问题 geometric consistency

⬅️ 返回 cs.CV 首页 · 🏠 返回主页