cs.CV（2024-03-01）

📊 共 15 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (5 🔗1) 支柱七：动作重定向 (Motion Retargeting) (3) 支柱三：空间感知与语义 (Perception & Semantics) (3 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (2) 支柱四：生成式动作 (Generative Motion) (2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (5 篇)

#题目一句话要点标签🔗⭐
1 Spurious Feature Eraser: Stabilizing Test-Time Adaptation for Vision-Language Foundation Model 提出Spurious Feature Eraser以解决视觉语言模型的决策捷径问题 foundation model
2 Exploring the dynamic interplay of cognitive load and emotional arousal by using multimodal measurements: Correlation of pupil diameter and emotional arousal in emotionally engaging tasks 通过多模态测量探讨认知负荷与情感唤起的动态关系 multimodal
3 Multimodal ArXiv: A Dataset for Improving Scientific Comprehension of Large Vision-Language Models 提出Multimodal ArXiv以提升科学理解能力的多模态数据集 multimodal
4 TempCompass: Do Video LLMs Really Understand Videos? 提出TempCompass基准以解决视频LLM时间感知能力不足问题 large language model ✅
5 Abductive Ego-View Accident Video Understanding for Safe Driving Perception 提出MM-AU数据集与AdVersa-SD框架以解决安全驾驶感知问题 multimodal

🔬 支柱七：动作重定向 (Motion Retargeting) (3 篇)

#题目一句话要点标签🔗⭐
6 MS-Net: A Multi-Path Sparse Model for Motion Prediction in Multi-Scenes 提出MS-Net以解决多场景下人类行为运动预测问题 motion prediction
7 Improving Explicit Spatial Relationships in Text-to-Image Generation through an Automatically Derived Dataset 提出自动生成数据集以改善文本到图像生成中的空间关系问题 spatial relationship
8 Can Transformers Capture Spatial Relations between Objects? 提出RelatiViT以解决空间关系预测问题 spatial relationship

🔬 支柱三：空间感知与语义 (Perception & Semantics) (3 篇)

#题目一句话要点标签🔗⭐
9 G3DR: Generative 3D Reconstruction in ImageNet 提出G3DR以解决单图像生成高质量3D对象的问题 3D reconstruction ✅
10 Multi-modal Attribute Prompting for Vision-Language Models 提出多模态属性提示方法以解决视觉语言模型的少样本问题 open-vocabulary open vocabulary
11 Trustworthy Self-Attention: Enabling the Network to Focus Only on the Most Relevant References 提出可信自注意力机制以解决光流预测中的遮挡问题 optical flow

🔬 支柱二：RL算法与架构 (RL & Architecture) (2 篇)

#题目一句话要点标签🔗⭐
12 Learning and Leveraging World Models in Visual Representation Learning 提出图像世界模型以扩展JEPA自监督学习能力 world model world models JEPA
13 Point Cloud Mamba: Point Cloud Learning via State Space Model 提出Mamba模型以高效处理点云数据 Mamba state space model

🔬 支柱四：生成式动作 (Generative Motion) (2 篇)

#题目一句话要点标签🔗⭐
14 Tri-Modal Motion Retrieval by Learning a Joint Embedding Space 提出LAVIMO框架以解决三模态运动检索问题 text-to-motion video-to-motion motion retrieval
15 CustomListener: Text-guided Responsive Interaction for User-friendly Listening Head Generation 提出CustomListener以解决用户自定义听众头生成问题 motion generation

⬅️ 返回 cs.CV 首页 · 🏠 返回主页