cs.CV（2024-03-13）

📊 共 27 篇论文 | 🔗 5 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (9) 支柱三：空间感知与语义 (Perception & Semantics) (8 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (4 🔗2) 支柱四：生成式动作 (Generative Motion) (2) 支柱六：视频提取与匹配 (Video Extraction) (2 🔗1) 支柱八：物理动画 (Physics-based Animation) (1) 支柱一：机器人控制 (Robot Control) (1 🔗1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (9 篇)

#题目一句话要点标签🔗⭐
1 CoIN: A Benchmark of Continual Instruction tuNing for Multimodel Large Language Model 提出CoIN基准以解决多模态大语言模型的持续指令调优问题 large language model multimodal instruction following
2 PathM3: A Multimodal Multi-Task Multiple Instance Learning Framework for Whole Slide Image Classification and Captioning 提出PathM3以解决全切片图像分类与描述对齐问题 multimodal
3 A Multimodal Fusion Network For Student Emotion Recognition Based on Transformer and Tensor Product 提出多模态融合网络以解决学生情感识别问题 multimodal
4 ALow-Cost Real-Time Framework for Industrial Action Recognition Using Foundation Models 提出低成本实时框架以解决工业动作识别问题 foundation model
5 SeCG: Semantic-Enhanced 3D Visual Grounding via Cross-modal Graph Attention 提出SeCG以解决3D视觉定位中的多对象识别问题 visual grounding
6 Masked Generative Story Transformer with Character Guidance and Caption Augmentation 提出基于掩码生成故事变换器以解决故事可视化问题 large language model
7 Unleashing the Power of Meta-tuning for Few-shot Generalization Through Sparse Interpolated Experts 提出稀疏元调优方法以解决少样本泛化问题 foundation model
8 Hierarchical Auto-Organizing System for Open-Ended Multi-Agent Navigation 提出层次自组织系统以解决多智能体导航问题 embodied AI
9 REPAIR: Rank Correlation and Noisy Pair Half-replacing with Memory for Noisy Correspondence 提出REPAIR框架以解决跨模态匹配中的噪声对应问题 multimodal

🔬 支柱三：空间感知与语义 (Perception & Semantics) (8 篇)

#题目一句话要点标签🔗⭐
10 GaussCtrl: Multi-View Consistent Text-Driven 3D Gaussian Splatting Editing 提出GaussCtrl以解决多视图一致性文本驱动3D场景编辑问题 3D gaussian splatting 3DGS gaussian splatting
11 Gaussian Splatting in Style 提出高效的3D场景风格化方法以解决多视角一致性问题 3D gaussian splatting gaussian splatting splatting
12 SM4Depth: Seamless Monocular Metric Depth Estimation across Multiple Cameras and Scenes by One Model 提出SM4Depth以解决单目度量深度估计的场景适应性问题 depth estimation metric depth foundation model ✅
13 METER: a mobile vision transformer architecture for monocular depth estimation 提出METER以解决嵌入式设备上的单目深度估计问题 depth estimation monocular depth
14 NeRF-Supervised Feature Point Detection and Description 提出基于NeRF的特征点检测与描述方法以提升视觉任务性能 visual SLAM NeRF neural radiance field
15 StyleDyRF: Zero-shot 4D Style Transfer for Dynamic Neural Radiance Fields 提出StyleDyRF以解决动态4D场景的风格迁移问题 NeRF neural radiance field
16 3DFIRES: Few Image 3D REconstruction for Scenes with Hidden Surface 提出3DFIRES以解决稀疏视图下的3D重建问题 3D reconstruction
17 SLCF-Net: Sequential LiDAR-Camera Fusion for Semantic Scene Completion using a 3D Recurrent U-Net 提出SLCF-Net以解决语义场景补全问题 Depth Anything

🔬 支柱二：RL算法与架构 (RL & Architecture) (4 篇)

#题目一句话要点标签🔗⭐
18 TINA: Think, Interaction, and Action Framework for Zero-Shot Vision Language Navigation 提出TINA框架以解决零-shot视觉语言导航问题 reinforcement learning vision-language-navigation VLN
19 FocusMAE: Gallbladder Cancer Detection from Ultrasound Videos with Focused Masked Autoencoders 提出FocusMAE以解决超声视频中的胆囊癌检测问题 representation learning masked autoencoder MAE ✅
20 MIM4D: Masked Modeling with Multi-View Video for Autonomous Driving Representation Learning 提出MIM4D以解决多视角视频的视觉表示学习问题 representation learning scene flow ✅
21 HIMap: HybrId Representation Learning for End-to-end Vectorized HD Map Construction 提出HIMap以解决高精度地图构建中的元素信息不足问题 representation learning

🔬 支柱四：生成式动作 (Generative Motion) (2 篇)

#题目一句话要点标签🔗⭐
22 Scaling Up Dynamic Human-Scene Interaction Modeling 提出TRUMANS数据集与新型HSI运动合成方法以解决数据稀缺问题 motion synthesis human-scene interaction HSI
23 VLOGGER: Multimodal Diffusion for Embodied Avatar Synthesis 提出VLOGGER以解决音频驱动的人类视频生成问题 motion diffusion model motion diffusion multimodal

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

#题目一句话要点标签🔗⭐
24 STMPL: Human Soft-Tissue Simulation 提出STMPL以解决人类软组织模拟的高效性问题 SMPL
25 Matching Semantically Similar Non-Identical Objects 提出语义增强加权方法以解决非同一物体匹配问题 feature matching ✅

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

#题目一句话要点标签🔗⭐
26 Spatiotemporal Diffusion Model with Paired Sampling for Accelerated Cardiac Cine MRI 提出时空扩散模型与配对采样以解决心脏Cine MRI重建模糊问题 spatiotemporal

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
27 Caltech Aerial RGB-Thermal Dataset in the Wild 提出RGB-热成像数据集以推动空中机器人感知算法发展 motion tracking ✅

⬅️ 返回 cs.CV 首页 · 🏠 返回主页