cs.CV（2025-05-31）

📊 共 15 篇论文 | 🔗 5 篇有代码

🎯 兴趣领域导航

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

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

#题目一句话要点标签🔗⭐
1 SatDreamer360: Multiview-Consistent Generation of Ground-Level Scenes from Satellite Imagery SatDreamer360：提出多视角一致的卫星图像到地面场景生成框架 dreamer height map
2 SenseFlow: Scaling Distribution Matching for Flow-based Text-to-Image Distillation SenseFlow：通过缩放分布匹配实现Flow模型文本到图像的蒸馏 flow matching distillation ✅
3 From Local Cues to Global Percepts: Emergent Gestalt Organization in Self-Supervised Vision Models 研究表明，自监督视觉模型通过Gestalt原则涌现全局感知能力，并提出DiSRT测试基准。 MAE spatial relationship
4 CReFT-CAD: Boosting Orthographic Projection Reasoning for CAD via Reinforcement Fine-Tuning CReFT-CAD：通过强化微调提升CAD正交投影推理能力 reinforcement learning instruction following

🔬 支柱一：机器人控制 (Robot Control) (3 篇)

#题目一句话要点标签🔗⭐
5 Multimodal Generative AI with Autoregressive LLMs for Human Motion Understanding and Generation: A Way Forward 综述：基于自回归LLM的多模态生成AI在人体运动理解与生成中的应用 humanoid text-to-motion motion synthesis
6 XYZ-IBD: A High-precision Bin-picking Dataset for Object 6D Pose Estimation Capturing Real-world Industrial Complexity 提出XYZ-IBD数据集，用于解决真实工业环境下物体6D位姿估计的难题。 manipulation depth estimation 6D pose estimation ✅
7 SEED: A Benchmark Dataset for Sequential Facial Attribute Editing with Diffusion Models 提出SEED数据集，用于评估扩散模型在人脸属性序列编辑中的性能，并提出FAITH模型。 manipulation ✅

🔬 支柱九：具身大模型 (Embodied Foundation Models) (3 篇)

#题目一句话要点标签🔗⭐
8 Chain-of-Frames: Advancing Video Understanding in Multimodal LLMs via Frame-Aware Reasoning 提出Chain-of-Frames，通过帧感知推理提升多模态LLM的视频理解能力 large language model multimodal chain-of-thought ✅
9 HueManity: Probing Fine-Grained Visual Perception in MLLMs HueManity：探究多模态大语言模型在细粒度视觉感知上的能力 large language model multimodal
10 Common Inpainted Objects In-N-Out of Context 提出COinCO数据集，用于提升模型对图像上下文一致性的理解和伪造检测能力。 large language model multimodal ✅

🔬 支柱三：空间感知与语义 (Perception & Semantics) (2 篇)

#题目一句话要点标签🔗⭐
11 Improving Optical Flow and Stereo Depth Estimation by Leveraging Uncertainty-Based Learning Difficulties 利用不确定性学习难度，提升光流和立体深度估计精度 depth estimation stereo depth optical flow
12 Test-time Vocabulary Adaptation for Language-driven Object Detection 提出VocAda，用于语言驱动目标检测的测试时词汇自适应，提升检测性能。 open-vocabulary open vocabulary

🔬 支柱六：视频提取与匹配 (Video Extraction) (1 篇)

#题目一句话要点标签🔗⭐
13 Sequence-Based Identification of First-Person Camera Wearers in Third-Person Views 提出基于序列的身份识别方法，用于在第三人称视角中识别第一人称相机佩戴者。 egocentric egocentric vision Ego4D

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
14 Parallel Rescaling: Rebalancing Consistency Guidance for Personalized Diffusion Models 提出并行重缩放方法，提升个性化扩散模型在少量样本下的prompt对齐度与图像质量 classifier-free guidance

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

#题目一句话要点标签🔗⭐
15 Event-based multi-view photogrammetry for high-dynamic, high-velocity target measurement 提出基于事件相机的多视图摄影测量方法，用于高动态高速目标测量。 spatiotemporal

⬅️ 返回 cs.CV 首页 · 🏠 返回主页