cs.CV（2026-02-06）

📊 共 30 篇论文 | 🔗 9 篇有代码

🎯 兴趣领域导航

支柱三：空间感知与语义 (Perception & Semantics) (7 🔗2) 支柱九：具身大模型 (Embodied Foundation Models) (7 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (6 🔗2) 支柱一：机器人控制 (Robot Control) (4) 支柱四：生成式动作 (Generative Motion) (3 🔗2) 支柱八：物理动画 (Physics-based Animation) (1 🔗1) 支柱七：动作重定向 (Motion Retargeting) (1) 支柱六：视频提取与匹配 (Video Extraction) (1)

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

#题目一句话要点标签🔗⭐
1 GaussianPOP: Principled Simplification Framework for Compact 3D Gaussian Splatting via Error Quantification GaussianPOP：基于误差量化的紧凑型3D高斯溅射简化框架 3D gaussian splatting 3DGS gaussian splatting
2 Forest canopy height estimation from satellite RGB imagery using large-scale airborne LiDAR-derived training data and monocular depth estimation 利用大规模机载LiDAR数据训练单目深度估计模型，实现卫星RGB影像森林冠层高度精确估计 depth estimation monocular depth Depth Anything
3 AdaptOVCD: Training-Free Open-Vocabulary Remote Sensing Change Detection via Adaptive Information Fusion 提出AdaptOVCD，一种免训练的遥感开放词汇变化检测方法，通过自适应信息融合实现。 open-vocabulary open vocabulary ✅
4 Uncertainty-Aware 4D Gaussian Splatting for Monocular Occluded Human Rendering 提出U-4DGS，解决单目视频中遮挡人体高保真渲染问题。 gaussian splatting splatting
5 Taming SAM3 in the Wild: A Concept Bank for Open-Vocabulary Segmentation 提出ConceptBank，解决SAM3在开放词汇分割中因数据/概念漂移导致的性能下降问题。 open-vocabulary open vocabulary ✅
6 POPL-KF: A Pose-Only Geometric Representation-Based Kalman Filter for Point-Line-Based Visual-Inertial Odometry 提出POPL-KF，一种基于位姿几何表示的点线视觉惯性里程计，提升复杂场景定位精度。 VIO optical flow motion estimation
7 CineScene: Implicit 3D as Effective Scene Representation for Cinematic Video Generation CineScene：利用隐式3D场景表示进行电影级视频生成 VGGT

🔬 支柱九：具身大模型 (Embodied Foundation Models) (7 篇)

#题目一句话要点标签🔗⭐
8 Towards Understanding Multimodal Fine-Tuning: Spatial Features 提出阶段式模型差异分析方法，揭示视觉-语言模型微调过程中空间特征的涌现。 multimodal visual grounding
9 Ex-Omni: Enabling 3D Facial Animation Generation for Omni-modal Large Language Models Ex-Omni：为全模态大语言模型实现语音驱动的3D面部动画生成 large language model multimodal
10 Prompt Reinjection: Alleviating Prompt Forgetting in Multimodal Diffusion Transformers 提出Prompt Reinjection，缓解多模态扩散Transformer中的Prompt遗忘问题 multimodal instruction following
11 Unsupervised MR-US Multimodal Image Registration with Multilevel Correlation Pyramidal Optimization 提出基于多层相关金字塔优化的无监督MR-US多模态图像配准方法 multimodal ✅
12 Zero-shot Multi-Contrast Brain MRI Registration by Intensity Randomizing T1-weighted MRI (LUMIR25) 提出基于强度随机化的零样本多对比度脑部MRI配准方法，提升跨域泛化性 foundation model multimodal
13 Seeing Beyond Redundancy: Task Complexity's Role in Vision Token Specialization in VLLMs 研究视觉语言模型中任务复杂度对视觉token特化的影响 large language model
14 Revisiting Salient Object Detection from an Observer-Centric Perspective 提出观察者中心显著性目标检测(OC-SOD)，解决传统方法主观性不足问题。 large language model ✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (6 篇)

#题目一句话要点标签🔗⭐
15 Zero-Shot UAV Navigation in Forests via Relightable 3D Gaussian Splatting 提出基于可重光照3D高斯溅射的零样本无人机森林导航方法 reinforcement learning 3D gaussian splatting gaussian splatting
16 DriveWorld-VLA: Unified Latent-Space World Modeling with Vision-Language-Action for Autonomous Driving DriveWorld-VLA：融合视觉-语言-动作的统一隐空间自动驾驶世界模型 world model vision-language-action VLA ✅
17 MedMO: Grounding and Understanding Multimodal Large Language Model for Medical Images MedMO：用于医学图像的具身和理解多模态大型语言模型 reinforcement learning large language model foundation model ✅
18 AS-Mamba: Asymmetric Self-Guided Mamba Decoupled Iterative Network for Metal Artifact Reduction 提出AS-Mamba，通过解耦迭代网络进行金属伪影抑制，提升CT图像质量。 Mamba SSM state space model
19 NanoFLUX: Distillation-Driven Compression of Large Text-to-Image Generation Models for Mobile Devices NanoFLUX：提出一种基于蒸馏的文本到图像生成模型压缩方法，用于移动设备。 flow matching distillation
20 SPARC: Separating Perception And Reasoning Circuits for Test-time Scaling of VLMs SPARC：分离视觉语言模型的感知与推理电路，实现测试时动态扩展。 reinforcement learning visual grounding

🔬 支柱一：机器人控制 (Robot Control) (4 篇)

#题目一句话要点标签🔗⭐
21 LIBERO-X: Robustness Litmus for Vision-Language-Action Models LIBERO-X：针对视觉-语言-动作模型的鲁棒性评测基准 manipulation teleoperation vision-language-action
22 MosaicThinker: On-Device Visual Spatial Reasoning for Embodied AI via Iterative Construction of Space Representation MosaicThinker：通过迭代构建空间表征，增强具身AI设备上的视觉空间推理能力 manipulation semantic map spatial relationship
23 CauCLIP: Bridging the Sim-to-Real Gap in Surgical Video Understanding via Causality-Inspired Vision-Language Modeling CauCLIP：通过因果视觉-语言建模弥合手术视频理解中的Sim-to-Real差距 sim-to-real
24 MicroBi-ConvLSTM: An Ultra-Lightweight Efficient Model for Human Activity Recognition on Resource Constrained Devices 提出MicroBi-ConvLSTM，一种用于资源受限设备上的人体活动识别超轻量级高效模型 locomotion

🔬 支柱四：生成式动作 (Generative Motion) (3 篇)

#题目一句话要点标签🔗⭐
25 Extended to Reality: Prompt Injection in 3D Environments 提出PI3D以解决3D环境中的提示注入攻击问题 physically plausible large language model multimodal
26 NECromancer: Breathing Life into Skeletons via BVH Animation 提出NECromancer以解决跨物种运动建模问题 motion tokenizer motion retrieval ✅
27 TLC-Plan: A Two-Level Codebook Based Network for End-to-End Vector Floorplan Generation TLC-Plan：一种基于双层码本网络的端到端矢量户型图生成方法 VQ-VAE ✅

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

#题目一句话要点标签🔗⭐
28 DuMeta++: Spatiotemporal Dual Meta-Learning for Generalizable Few-Shot Brain Tissue Segmentation Across Diverse Ages DuMeta++：时空双重元学习用于跨年龄泛化的少样本脑组织分割 spatiotemporal ✅

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

#题目一句话要点标签🔗⭐
29 An Integer Linear Programming Approach to Geometrically Consistent Partial-Partial Shape Matching 提出一种基于整数线性规划的几何一致性部分-部分三维形状匹配方法 geometric consistency

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

#题目一句话要点标签🔗⭐
30 Bridging the Indoor-Outdoor Gap: Vision-Centric Instruction-Guided Embodied Navigation for the Last Meters 提出视觉中心指令引导的室内外无先验知识具身导航方法 egocentric

⬅️ 返回 cs.CV 首页 · 🏠 返回主页