cs.CV（2026-11-02）

📊 共 47 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) (17 🔗5) 支柱二：RL算法与架构 (RL & Architecture) (11 🔗3) 支柱三：空间感知与语义 (Perception & Semantics) (5 🔗1) 支柱七：动作重定向 (Motion Retargeting) (4 🔗1) 支柱一：机器人控制 (Robot Control) (4) 支柱六：视频提取与匹配 (Video Extraction) (3 🔗1) 支柱八：物理动画 (Physics-based Animation) (2) 支柱四：生成式动作 (Generative Motion) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (17 篇)

#题目一句话要点标签🔗⭐
1 Q-CamTuning: A Multimodal Large Language Model for Fine-Grained Quality Reasoning in Livestreaming Camera Tuning 提出Q-CamTuning以解决直播摄像头调优中的质量推理问题 large language model multimodal
2 Lookahead Prediction-Guided Adaptive Visual Refocusing for Interleaved Multimodal Reasoning 提出LPAVR以解决多模态推理中的时间证据近视问题 large language model multimodal chain-of-thought
3 Latent Spectral Prompting for Multimodal Hyperspectral Object Tracking 提出SProTrack以解决多模态高光谱目标跟踪问题 foundation model multimodal
4 Blueprint2CAD: Benchmarking Foundation Models on Engineering Drawing-to-CAD Render Correspondence 提出Blueprint2CAD以解决工程图与CAD渲染对应问题 foundation model chain-of-thought
5 Calibrating Multimodal Test-Time Adaptation under Reliability and Availability Modality Shifts 提出AdaPGC以解决多模态测试时适应中的可靠性与可用性问题 multimodal ✅
6 Multimodal Prompt Learning with Dual-Granularity Contrastive Alignment for Preoperative Prediction of Lung Adenocarcinoma Invasiveness Using 3D CT 提出多模态提示学习框架以解决肺腺癌侵袭性预测问题 multimodal
7 VL-Dedup: Graded Multimodal Entity Resolution for Image-Text Data Deduplication 提出VL-Dedup以解决图像-文本数据去重问题 multimodal
8 Unlocking the Potential of Vision Foundation Models for Multi-Task Dense Prediction via Mixture of Task Adapters 提出D2MoE以解决多任务密集预测中的任务适配问题 foundation model
9 Probe-Guided Gradient Balancing for Multimodal Learning 提出探针引导的梯度平衡方法以解决多模态学习中的不平衡问题 multimodal ✅
10 ReinSeg: Retrieval-Guided Bounded Multimodal Refinement for 3-D Road Anomaly Segmentation 提出ReinSeg以解决3D道路异常分割问题 multimodal
11 Reviso: Not All Visual Evidence Deserves Equal Attention in Video Captioning 提出Reviso框架以优化视频字幕生成中的视觉信息关注 large language model multimodal
12 Align Once to Explain: Feature Alignment for Scalable B-cosification of Foundational Vision Transformers 提出ALOE以提高基础视觉模型的可解释性 zero-shot transfer
13 DR Pixel: Page-Native Deep Research with Vision-Language Models 提出DR Pixel以解决复杂文档深度研究问题 multimodal
14 ProtoFusion: Protocol-Guided Prompt Refinement for Generative Infrared and Visible Image Fusion 提出ProtoFusion以解决红外与可见光图像融合问题 multimodal ✅
15 SeMo3D: Unifying 3D Generation, Editing, and Completion via Selective Modeling 提出SeMo3D以解决3D生成、编辑和补全任务的统一问题 multimodal ✅
16 PixelBench: Can LLMs Understand and Generate Images Pixel by Pixel? 提出PixelBench以评估LLMs在像素生成中的能力 large language model
17 Exploiting Textual Semantics for Robust Cross-View Object Correspondence 提出TeSCo框架以解决复杂场景下的跨视角物体对应问题 multimodal ✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (11 篇)

#题目一句话要点标签🔗⭐
18 SurgPace: Complexity-Aware Progressive Distillation for Surgical Multimodal Large Language Models 提出SurgPace以解决外科多模态大语言模型的蒸馏挑战 teacher-student distillation scene understanding ✅
19 Fuse To Align: Multimodal Representation Learning for Retrieval via Hypergraph Regularization 提出超图融合框架以解决多模态检索中的信息缺失问题 representation learning multimodal
20 MIRA: Reinforcing Multimodal Reasoning via Deceptive Contextual Augmentation 提出MIRA以解决多模态推理中的感知-推理差距问题 reinforcement learning multimodal
21 History Forcing: Distilling History-Dependent Continuations for Long-Horizon Video World Models 提出历史强制方法以解决长视频生成中的一致性问题 world model world models distillation
22 MirrorComm: A Communication-Efficient Cooperative Perception Framework via Mirrored World Models 提出MirrorComm以解决协同感知中的通信效率问题 world model world models
23 OmniOPD: Stable On-Policy Distillation for Joint Audio-Video Generation 提出OmniOPD以解决音视频生成中的跨模态干扰问题 reinforcement learning teacher-student distillation
24 Contrastive Adapter Training and Consensus Knowledge Distillation for Multi-Source-Free Domain Adaptation in Skin Cancer Diagnosis 提出CAT-CKD以解决皮肤癌诊断中的多源无监督领域适应问题 contrastive learning distillation ✅
25 HIVE-Diff: High Frame Rate HDR Video Reconstruction from a Single LDR Frame and Events 提出HIVE-Diff以解决单帧LDR和事件重建高帧率HDR视频问题 teacher-student distillation TAMP
26 Efficient Hybrid Distillation: Synergizing Score and Adversarial Objectives for One-Step Diffusion 提出高效混合蒸馏方法以优化扩散模型训练 flow matching distillation ✅
27 DINOFAS: Bonafide-Centric Representation Learning for Face Anti-Spoofing 提出DINOFAS框架以解决面部反欺诈中的数据集特定问题 representation learning
28 Isekai3D SDG: A Platform for Diverse Synthetic Gameplay Visual Data 提出Isekai3D SDG以生成多样化的合成游戏视觉数据 world model world models

🔬 支柱三：空间感知与语义 (Perception & Semantics) (5 篇)

#题目一句话要点标签🔗⭐
29 OctreeSplat: Hierarchical Octree-Guided Feed-Forward Gaussian Splatting 提出OctreeSplat以解决3D重建中的高效高质量问题 3D gaussian splatting 3DGS 3D reconstruction
30 3D Gaussian Splatting within Self-Learned Neural View-Dependent Color Fields 提出自学习神经视图依赖颜色场以解决3D高斯点云建模问题 3D gaussian splatting 3DGS gaussian splatting ✅
31 GaussFlaw: Fine-Grained Perceptual Diagnosis for 3D Gaussian Splatting 提出GaussFlaw以解决3D Gaussian Splatting的细粒度诊断问题 3D gaussian splatting 3DGS gaussian splatting
32 CoLa3D: Composable Latent 3D Decomposition 提出CoLa3D以解决3D场景可组合分解问题 sam 3D SAM 3D
33 AnchorComplete: Recovering Complete Scene Meshes from Partial Point Clouds 提出AnchorComplete以解决部分点云重建完整场景网格的问题 scene reconstruction

🔬 支柱七：动作重定向 (Motion Retargeting) (4 篇)

#题目一句话要点标签🔗⭐
34 Trace4D: Grounded Spatiotemporal Reasoning via 4D State Sequences 提出Trace4D以解决时空推理中的4D状态演变问题 spatial relationship spatiotemporal chain-of-thought
35 Solv3R: 3D Looped Transformer as a Multi-view Solver 提出Solv3R以解决多视角几何重建问题 motion reconstruction foundation model ✅
36 MoSoLoco: Monocular 3D Human Localization through Visual Latent Features and Social Context 提出MoSoLoco以解决单目3D人类定位问题 motion prediction
37 UniCross3D: Unified Cross-View and Cross-Domain Diffusion for Consistent Single-Image 3D Generation 提出UniCross3D以解决单图像3D生成中的一致性问题 geometric consistency

🔬 支柱一：机器人控制 (Robot Control) (4 篇)

#题目一句话要点标签🔗⭐
38 Evaluating Depth-Augmented Visual Features for Cross-Dataset Lunar Traversability Estimation 提出弱监督框架以解决月球表面可 traversability 估计问题 teleoperation traversability multimodal
39 EgoActionBench: Benchmarking Dense Action Captioning for Egocentric Human Videos 提出EgoActionBench以解决自我中心视频中的动作理解问题 manipulation reinforcement learning egocentric
40 TailorRig: Controllable and Structured Auxiliary Skeleton Generation for Garment-Aware Rigging 提出TailorRig以解决复杂服装人形角色的辅助骨架生成问题 humanoid
41 3L-D: A Low-Light Level Dataset from a Sensor-Agnostic, Alignment-Preserving Acquisition Framework 提出3L-D数据集以解决低光照图像增强的现有数据集局限性 manipulation

🔬 支柱六：视频提取与匹配 (Video Extraction) (3 篇)

#题目一句话要点标签🔗⭐
42 Mesh-Agnostic Refinement Models for Latent Space Encoding of 3D Human Body Shapes 提出网格无关的精炼模型以改进3D人体形状的潜在空间编码 SMPL
43 MuCHMDPoser: Multiview RGB-Camera Full-Body Pose Tracking for Virtual Reality 提出MuCHMDPoser以解决虚拟现实中的全身姿态跟踪问题 sparse sensors ✅
44 GEAR: Generative Editing for Scene-Conditioned Outfit Retrieval for VR Travel 提出GEAR以解决虚拟环境中服装检索的场景适配问题 egocentric

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

#题目一句话要点标签🔗⭐
45 R$^2$-Omni: Reducing Structural and Precision Redundancy in Omni-LLMs for Long-Form Generation 提出R$^2$-Omni以解决Omni-LLMs在长文本生成中的冗余问题 spatiotemporal large language model multimodal
46 MoviEdit: Long-form Multi-Shot Video Editing Made Easy 提出MoviEdit以解决长视频多镜头编辑的挑战 spatiotemporal

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
47 RouteREPA: Routed Representation Alignment for Just Image Transformers 提出RouteREPA以解决Just Image Transformers的表示对齐问题 classifier-free guidance

⬅️ 返回 cs.CV 首页 · 🏠 返回主页