cs.CV（2023-11-30）

📊 共 58 篇论文 | 🔗 14 篇有代码

🎯 兴趣领域导航

支柱三：空间感知与语义 (Perception & Semantics) (18 🔗1) 支柱九：具身大模型 (Embodied Foundation Models) (16 🔗7) 支柱二：RL算法与架构 (RL & Architecture) (12 🔗4) 支柱一：机器人控制 (Robot Control) (4) 支柱四：生成式动作 (Generative Motion) (3 🔗1) 支柱六：视频提取与匹配 (Video Extraction) (2) 支柱八：物理动画 (Physics-based Animation) (2 🔗1) 支柱七：动作重定向 (Motion Retargeting) (1)

🔬 支柱三：空间感知与语义 (Perception & Semantics) (18 篇)

#题目一句话要点标签🔗⭐
1 CompGS: Smaller and Faster Gaussian Splatting with Vector Quantization 提出CompGS以解决3D高斯点云存储与渲染效率问题 3D gaussian splatting 3DGS gaussian splatting
2 DynMF: Neural Motion Factorization for Real-time Dynamic View Synthesis with 3D Gaussian Splatting 提出DynMF以解决动态场景合成中的运动建模问题 3D gaussian splatting gaussian splatting splatting
3 Language Embedded 3D Gaussians for Open-Vocabulary Scene Understanding 提出语言嵌入3D高斯以解决开放词汇场景理解问题 scene understanding open-vocabulary open vocabulary
4 SparseGS: Real-Time 360° Sparse View Synthesis using Gaussian Splatting 提出SparseGS以解决稀疏视图合成问题 3D gaussian splatting 3DGS gaussian splatting
5 PyNeRF: Pyramidal Neural Radiance Fields 提出PyNeRF以解决NeRF渲染中的缩放问题 NeRF neural radiance field
6 Multi-task learning with cross-task consistency for improved depth estimation in colonoscopy 提出多任务学习方法以改善结肠镜下深度估计 depth estimation 3D reconstruction
7 Redefining Recon: Bridging Gaps with UAVs, 360 degree Cameras, and Neural Radiance Fields 提出一种新方法以利用无人机和NeRF生成灾后3D模型 NeRF neural radiance field
8 ZeST-NeRF: Using temporal aggregation for Zero-Shot Temporal NeRFs 提出ZeST-NeRF以解决零-shot时间NeRF生成问题 NeRF scene flow
9 HOLD: Category-agnostic 3D Reconstruction of Interacting Hands and Objects from Video 提出HOLD方法以解决无类别限制的3D手与物体重建问题 3D reconstruction hand-object reconstruction ✅
10 Scaffold-GS: Structured 3D Gaussians for View-Adaptive Rendering 提出Scaffold-GS以解决3D高斯渲染冗余问题 3D gaussian splatting gaussian splatting splatting
11 FoundPose: Unseen Object Pose Estimation with Foundation Features 提出FoundPose以解决未知物体姿态估计问题 6D pose estimation foundation model
12 Un-EVIMO: Unsupervised Event-Based Independent Motion Segmentation 提出无监督事件驱动的独立运动分割方法以解决标注依赖问题 optical flow IMoS
13 DNS SLAM: Dense Neural Semantic-Informed SLAM 提出DNS SLAM以解决复杂场景下的重建平滑问题 implicit representation
14 Portrait4D: Learning One-Shot 4D Head Avatar Synthesis using Synthetic Data 提出一种基于合成数据的一次性4D头部头像合成方法 3D reconstruction
15 DiffCAD: Weakly-Supervised Probabilistic CAD Model Retrieval and Alignment from an RGB Image 提出DiffCAD以解决CAD模型检索与对齐问题 monocular depth
16 MRFP: Learning Generalizable Semantic Segmentation from Sim-2-Real with Multi-Resolution Feature Perturbation 提出MRFP以解决模拟到真实语义分割的泛化问题 scene understanding
17 CosAvatar: Consistent and Animatable Portrait Video Tuning with Text Prompt 提出CosAvatar以解决数字肖像视频编辑中的一致性问题 NeRF
18 SimulFlow: Simultaneously Extracting Feature and Identifying Target for Unsupervised Video Object Segmentation 提出SimulFlow以解决无监督视频目标分割中的特征提取与目标识别问题 optical flow

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

#题目一句话要点标签🔗⭐
19 Merlin:Empowering Multimodal LLMs with Foresight Minds 提出未来建模方法以增强多模态大语言模型的推理能力 large language model multimodal
20 BioCLIP: A Vision Foundation Model for the Tree of Life 提出BioCLIP以解决生物图像分类的通用性问题 foundation model ✅
21 CoDi-2: In-Context, Interleaved, and Interactive Any-to-Any Generation 提出CoDi-2以解决复杂多模态指令理解问题 large language model foundation model multimodal
22 Spacewalk-18: A Benchmark for Multimodal and Long-form Procedural Video Understanding in Novel Domains 提出Spacewalk-18基准以解决多模态长视频理解问题 multimodal ✅
23 Which way is `right'?: Uncovering limitations of Vision-and-Language Navigation model 提出掩蔽实验以揭示视觉与语言导航模型的局限性 VLN
24 TrafficMOT: A Challenging Dataset for Multi-Object Tracking in Complex Traffic Scenarios 提出TrafficMOT数据集以解决复杂交通场景中的多目标跟踪问题 foundation model
25 VIDiff: Translating Videos via Multi-Modal Instructions with Diffusion Models 提出VIDiff以解决视频编辑与理解的多模态指令问题 foundation model ✅
26 InstructSeq: Unifying Vision Tasks with Instruction-conditioned Multi-modal Sequence Generation 提出InstructSeq框架以统一多种视觉任务 multimodal ✅
27 A Video is Worth 10,000 Words: Training and Benchmarking with Diverse Captions for Better Long Video Retrieval 提出多样化字幕生成方法以提升长视频检索性能 large language model ✅
28 X-InstructBLIP: A Framework for aligning X-Modal instruction-aware representations to LLMs and Emergent Cross-modal Reasoning 提出X-InstructBLIP框架以解决跨模态推理问题 large language model
29 CLIP-QDA: An Explainable Concept Bottleneck Model 提出CLIP-QDA以实现快速可解释的图像分类 foundation model
30 Detailed Human-Centric Text Description-Driven Large Scene Synthesis 提出DetText2Scene以解决文本驱动大场景合成的控制问题 large language model
31 Semantic-Aware Frame-Event Fusion based Pattern Recognition via Large Vision-Language Models 提出基于语义感知的帧-事件融合模式识别方法以解决现有技术的不足 multimodal ✅
32 OST: Refining Text Knowledge with Optimal Spatio-Temporal Descriptor for General Video Recognition 提出OST方法以优化视频识别中的文本知识 large language model
33 VTimeLLM: Empower LLM to Grasp Video Moments 提出VTimeLLM以解决视频时序理解问题 large language model
34 TLDR: Text Based Last-layer Retraining for Debiasing Image Classifiers 提出TLDR以解决图像分类器偏见问题 large language model ✅

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

#题目一句话要点标签🔗⭐
35 Periodic Vibration Gaussian: Dynamic Urban Scene Reconstruction and Real-time Rendering 提出周期振动高斯模型以解决动态城市场景重建问题 representation learning 3D gaussian splatting gaussian splatting
36 Anisotropic Neural Representation Learning for High-Quality Neural Rendering 提出各向异性神经表示学习以提升神经渲染质量 representation learning NeRF neural radiance field
37 Knowledge Transfer from Vision Foundation Models for Efficient Training of Small Task-specific Models 提出任务导向知识转移方法以高效训练小型任务特定模型 distillation foundation model
38 MLLMs-Augmented Visual-Language Representation Learning 提出MLLMs增强视觉语言表示学习以提升图像文本检索性能 representation learning large language model
39 GraphDreamer: Compositional 3D Scene Synthesis from Scene Graphs 提出GraphDreamer以解决复杂场景3D建模问题 dreamer penetration
40 Contrastive Denoising Score for Text-guided Latent Diffusion Image Editing 提出对比去噪评分以解决图像编辑中的结构保留问题 contrastive learning distillation NeRF ✅
41 LucidDreaming: Controllable Object-Centric 3D Generation 提出LucidDreaming以解决3D生成控制不足问题 distillation NeRF large language model
42 Lasagna: Layered Score Distillation for Disentangled Object Relighting 提出Lasagna以解决图像重光照控制问题 distillation
43 Dataset Distillation via Curriculum Data Synthesis in Large Data Era 通过课程数据合成提出数据集蒸馏方法以提高训练效率 distillation ✅
44 Dataset Distillation via the Wasserstein Metric 提出WMDD方法以解决数据集蒸馏效率问题 distillation
45 E2PNet: Event to Point Cloud Registration with Spatio-Temporal Representation Learning 提出E2PNet以解决事件相机与点云注册问题 representation learning ✅
46 X-Dreamer: Creating High-quality 3D Content by Bridging the Domain Gap Between Text-to-2D and Text-to-3D Generation 提出X-Dreamer以解决文本到3D内容生成中的领域差距问题 dreamer ✅

🔬 支柱一：机器人控制 (Robot Control) (4 篇)

#题目一句话要点标签🔗⭐
47 DeformGS: Scene Flow in Highly Deformable Scenes for Deformable Object Manipulation 提出DeformGS以解决高度可变形场景中的场景流问题 manipulation real2sim imitation learning
48 Towards Unsupervised Representation Learning: Learning, Evaluating and Transferring Visual Representations 提出无监督视觉表征学习方法以提升模型泛化能力 sim-to-real representation learning
49 Learning Part Segmentation from Synthetic Animals 提出CB-FDM方法以解决合成动物部件分割的领域适应问题 quadruped
50 Advances in 3D Neural Stylization: A Survey 提出神经网络驱动的3D艺术风格化方法以解决创作挑战 manipulation

🔬 支柱四：生成式动作 (Generative Motion) (3 篇)

#题目一句话要点标签🔗⭐
51 OmniMotionGPT: Animal Motion Generation with Limited Data 提出OmniMotionGPT以解决动物运动生成数据不足问题 motion synthesis motion generation MotionGPT
52 Probabilistic Speech-Driven 3D Facial Motion Synthesis: New Benchmarks, Methods, and Applications 提出概率模型以解决语音驱动的3D面部动画问题 motion synthesis
53 On Exact Inversion of DPM-Solvers 提出精确反演算法以解决DPM求解器的逆问题 classifier-free guidance ✅

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

#题目一句话要点标签🔗⭐
54 ChatPose: Chatting about 3D Human Pose 提出ChatPose以解决3D人类姿态理解与生成问题 SMPL large language model multimodal
55 Ego-Exo4D: Understanding Skilled Human Activity from First- and Third-Person Perspectives 提出Ego-Exo4D数据集以解决人类活动理解的多视角挑战 egocentric multimodal

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

#题目一句话要点标签🔗⭐
56 AV-RIR: Audio-Visual Room Impulse Response Estimation 提出AV-RIR以解决房间脉冲响应估计问题 PULSE
57 RainAI -- Precipitation Nowcasting from Satellite Data 提出RainAI以解决卫星数据降水预测问题 spatiotemporal ✅

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

#题目一句话要点标签🔗⭐
58 Fixed-length Dense Descriptor for Efficient Fingerprint Matching 提出固定长度密集描述符以解决指纹匹配效率问题 spatial relationship

⬅️ 返回 cs.CV 首页 · 🏠 返回主页