cs.CV（2026-09-24）

📊 共 43 篇论文 | 🔗 11 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (18 🔗3) 支柱三：空间感知与语义 (Perception & Semantics) (10 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (9 🔗4) 支柱六：视频提取与匹配 (Video Extraction) (2 🔗2) 支柱一：机器人控制 (Robot Control) (2) 支柱七：动作重定向 (Motion Retargeting) (1) 支柱八：物理动画 (Physics-based Animation) (1)

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

#题目一句话要点标签🔗⭐
1 The Alignment Illusion in Multimodal Large Language Models 提出主成分角度差距以解决多模态大语言模型的对齐幻觉问题 large language model multimodal
2 Retrieve-to-Localize: Bridging Large Language Models and LiDAR Geometry for Spatial Grounding 提出Retrieve-to-Localize以解决空间定位问题 large language model language conditioned
3 Multimodal Thinking with Renderable Programs 提出SVGLM框架以解决视觉语言模型的图像推理问题 multimodal
4 Mind What Matters for Reasoning: Aligning Cross-Modal Attention via Selective Probability Mass Concentration 提出选择性概率质量集中方法以增强多模态推理能力 large language model multimodal visual grounding
5 A Multimodal Dataset for Survival Prediction in Resected Pancreatic Ductal Adenocarcinoma 提出多模态数据集以改善胰腺导管腺癌生存预测 multimodal
6 Hyperbolic Multimodal Continual Learning: A Closest-Admissible Solution 提出超曲面多模态持续学习方法以解决几何保持问题 multimodal
7 Multimodal Routing and Region Refinement for Language-Guided Medical Image Segmentation 提出MRSeg以解决医学图像分割中的模糊性问题 multimodal ✅
8 Industrial Anomaly Detection via Defect-Grounded Reasoning in Visual Latent Space 提出Anomaly-LR以解决工业异常检测中的缺陷推理问题 large language model multimodal ✅
9 MoVISA: Multi-Token Reasoning for Video Object Segmentation 提出MoVISA以解决视频物体分割中的单一标记局限性问题 large language model multimodal
10 Exploiting Target Knowledge from MLLMs for Robust Few-Shot Segmentation 提出MK-FSS框架以解决少样本分割中的目标知识不足问题 large language model multimodal
11 Training-Free Hold-Usage Detection in Sport Climbing with Foundation Pose Models 提出无训练的攀岩抓握使用检测方法以解决现有模型不足问题 foundation model
12 GHOST-Q: Towards Studying Grounding Hallucinations Overlooked Under Same-score TradeOffs in Quantized VLMS 提出GHOST-Q以解决量化视觉语言模型中的幻觉问题 visual grounding
13 SEE Challenge 2026: Event-Guided Brightness Adjustment Across a Broad Illumination Range 提出事件引导的亮度调整方法以解决广泛照明范围下的图像恢复问题 multimodal
14 Recoverable Geographic Location Information in Earth-Observation Embeddings 评估地理信息在地球观测嵌入中的可恢复性 foundation model
15 RGBD20K: A Large-Scale Benchmark for RGB-D Semantic Segmentation 提出RGBD20K数据集以推动RGB-D语义分割研究 multimodal ✅
16 What Improves Multimodal Misinformation Detection? Answers from a Large-Scale Empirical Study 通过大规模实证研究提升多模态虚假信息检测效果 multimodal
17 ProCAP: Probabilistic Cross-Attentive Prompt Learning for Vision-Language Models 提出ProCAP以解决视觉-语言模型在低样本学习中的脆弱性问题 multimodal
18 AlphaEarth distinguishes cities but compresses urban variation 提出AlphaEarth以解决城市比较中的数据压缩问题 foundation model

🔬 支柱三：空间感知与语义 (Perception & Semantics) (10 篇)

#题目一句话要点标签🔗⭐
19 OceanXL: Large-scale Underwater 3D Gaussian Splatting via Block Partitioning and Adaptive Pruning 提出OceanXL以解决大规模水下3D重建问题 3D gaussian splatting 3DGS 3D reconstruction
20 Towards Practical Compression of 3D Gaussian Splatting 提出COSA-GS以解决3D高斯点云压缩问题 3D gaussian splatting 3DGS gaussian splatting ✅
21 Only What Was Seen: Observation-Gram Compaction of View-Dependent Appearance in 3D Gaussian Splatting 提出观察Gram矩阵压缩方法以优化3D高斯点云表现 3D gaussian splatting gaussian splatting splatting
22 SplatLabel: Pseudo-Labelling through 4D Gaussian Splatting 提出SplatLabel以解决3D语义伪标注问题 gaussian splatting splatting occupancy grid
23 WildHSR: Metric Feed-Forward 4D People-Scene Reconstruction from a 3D Foundation Model 提出WildHSR以解决视频中人场景重建问题 scene reconstruction foundation model
24 ADATEX4D: adaptive texture capacity allocation for 4D gaussian splatting 提出AdaTex4D以解决4D高斯点云纹理分配不均问题 gaussian splatting splatting
25 PlenoCI: Plenoptic CharacterIstics for View Dependence Aware Change Classification 提出PlenoCI以解决视依赖变化分类问题 3D gaussian splatting 3DGS gaussian splatting ✅
26 FounRef: Robust, Structure-Preserving, and Fast Metric Refinement of Frozen Monocular Foundation Priors with Sparse Anchors 提出FounRef以解决单目基础模型的深度度量问题 metric depth MoGe foundation model
27 Seeing Is Not Measuring: Tool-Augmented Metric Spatial Reasoning for Vision-Language Models 提出工具增强的度量空间推理框架以提升视觉语言模型性能 depth estimation metric depth egocentric
28 LiTe-GS: Oracle-Efficient Next Best View Selection for 3D Gaussian Splatting 提出LiTe-GS以解决3D高斯点云视角选择效率问题 3D gaussian splatting gaussian splatting splatting

🔬 支柱二：RL算法与架构 (RL & Architecture) (9 篇)

#题目一句话要点标签🔗⭐
29 IronViT: Toward Efficient Generalist Visual Representation Learning 提出IronViT以解决高分辨率视觉表示学习的计算成本问题 linear attention representation learning distillation
30 AV-GRPO: Modality-Anchored Decoupling Diffusion Reinforcement Learning for Joint Audio-Video Generation 提出AV-GRPO以解决音视频生成中的模态对齐与同步问题 reinforcement learning preference learning multimodal ✅
31 HelloWorld: Towards Practical Applications of Generative Driving World Models 提出HelloWorld以解决可扩展驾驶数据生成问题 world model world models
32 ComplexSync: High-Fidelity and Real-Time Lip Sync in Complex Scenarios 提出ComplexSync以解决复杂场景下的高保真实时唇动同步问题 distillation foundation model
33 Spectral Amplitude Purification in Distribution Matching for Diffusion Distillation 提出谱幅度净化方法以解决扩散蒸馏中的分布匹配问题 distillation
34 A Hybrid CNN--State-Space--Attention Backbone with Joint-Embedding Predictive Pretraining for 12-Lead ECG Classification 提出混合CNN-状态空间-注意力骨干网络以解决12导联ECG分类问题 JEPA SSM ✅
35 Action Forcing: Training World Models on Unsupervised Video by Recovering Underlying Egomotion Bases 提出基于无监督视频的动作强制训练世界模型方法 world model world models
36 AV-GRPO: Modality-Anchored Decoupling Diffusion Reinforcement Learning for Joint Audio-Video Generation 提出AV-GRPO以解决音视频生成中的模态同步问题 reinforcement learning preference learning multimodal ✅
37 The Shape of Events: Edge-Based Inductive Biases via Cross-Domain Distillation 通过跨域蒸馏提出边缘基础的归纳偏置以解决视觉模型的鲁棒性问题 distillation ✅

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

#题目一句话要点标签🔗⭐
38 Ego-Exo4D Human Meshes Dataset: 4D Human Motion Reconstruction for Ego-Exo Captures 提出Ego-Exo4D-HM以解决稀疏3D人类姿态重建问题 egocentric human motion human motion reconstruction ✅
39 PHOSA: Photorealistic 3D Sign Avatar Modeling and Benchmark 提出PHOSA以解决手语虚拟化建模问题 SMPL SMPL-X ✅

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
40 Direction-Scale Decomposition in Action Representation: Rethinking What to Tokenize for Vision-Language-Action Models 提出方向-尺度分解以改善视觉-语言-动作模型的动作表示 manipulation vision-language-action VLA
41 EIB-Net: Entropy-Guided Information Bottleneck for Generalizable AI-Generated Image Detection 提出EIB-Net以解决AI生成图像检测的挑战 manipulation

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

#题目一句话要点标签🔗⭐
42 FluidRain: Incompressible Rain Flow as an Attention Bias for Loop-in-Loop Video Deraining 提出FluidRain以解决视频去雨中的运动估计不可靠问题 motion estimation spatiotemporal

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

#题目一句话要点标签🔗⭐
43 Shadow Reduction in Ultrasound Imaging Using Differentiable Simulation and Radiance Field Decomposition 提出RFlash以解决超声成像中的阴影问题 differentiable simulation

⬅️ 返回 cs.CV 首页 · 🏠 返回主页