cs.CV（2025-07-28）

📊 共 29 篇论文 | 🔗 8 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (13 🔗6) 支柱三：空间感知与语义 (Perception & Semantics) (5 🔗1) 支柱一：机器人控制 (Robot Control) (4) 支柱二：RL算法与架构 (RL & Architecture) (4 🔗1) 支柱七：动作重定向 (Motion Retargeting) (2) 支柱四：生成式动作 (Generative Motion) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (13 篇)

#题目一句话要点标签🔗⭐
1 Multimodal LLMs as Customized Reward Models for Text-to-Image Generation 提出LLaVA-Reward，利用多模态LLM为文本到图像生成定制奖励模型 large language model multimodal instruction following
2 Not Only Grey Matter: OmniBrain for Robust Multimodal Classification of Alzheimer's Disease OmniBrain：用于阿尔茨海默病多模态稳健分类的统一框架 multimodal
3 RingMo-Agent: A Unified Remote Sensing Foundation Model for Multi-Platform and Multi-Modal Reasoning 提出RingMo-Agent，用于多平台多模态遥感图像的统一推理。 foundation model
4 ATR-UMMIM: A Benchmark Dataset for UAV-Based Multimodal Image Registration under Complex Imaging Conditions ATR-UMMIM：无人机多模态图像配准基准数据集，应对复杂成像条件 multimodal ✅
5 A Multimodal Architecture for Endpoint Position Prediction in Team-based Multiplayer Games 提出一种多模态架构，用于预测团队多人游戏中玩家的未来位置。 multimodal
6 GPT-IMAGE-EDIT-1.5M: A Million-Scale, GPT-Generated Image Dataset 提出GPT-IMAGE-EDIT-1.5M大规模图像编辑数据集，促进开源指令引导图像编辑研究。 multimodal instruction following
7 Security Tensors as a Cross-Modal Bridge: Extending Text-Aligned Safety to Vision in LVLM 提出安全张量，将文本对齐的安全性扩展到LVLM中的视觉模态 large language model multimodal
8 T2I-Copilot: A Training-Free Multi-Agent Text-to-Image System for Enhanced Prompt Interpretation and Interactive Generation 提出T2I-Copilot，一种无需训练的多智能体文本到图像系统，增强提示理解和交互式生成。 large language model multimodal ✅
9 On Explaining Visual Captioning with Hybrid Markov Logic Networks 提出基于混合马尔可夫逻辑网络的视觉描述解释框架，提升模型可解释性。 multimodal
10 METEOR: Multi-Encoder Collaborative Token Pruning for Efficient Vision Language Models 提出METEOR，通过多编码器协同Token剪枝提升视觉语言模型效率 multimodal ✅
11 KASportsFormer: Kinematic Anatomy Enhanced Transformer for 3D Human Pose Estimation on Short Sports Scene Video KASportsFormer：运动解剖学增强Transformer，用于短视频运动场景3D人体姿态估计 multimodal ✅
12 TransPrune: Token Transition Pruning for Efficient Large Vision-Language Model TransPrune：面向高效大型视觉-语言模型的Token转移剪枝方法 multimodal ✅
13 T2VParser: Adaptive Decomposition Tokens for Partial Alignment in Text to Video Retrieval 提出T2VParser，通过自适应分解token实现文本到视频检索中的局部对齐。 multimodal ✅

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

#题目一句话要点标签🔗⭐
14 $S^3$LAM: Surfel Splatting SLAM for Geometrically Accurate Tracking and Mapping S³LAM：提出基于Surfel Splatting的SLAM系统，实现高精度几何跟踪与建图 3DGS splatting
15 Endoscopic Depth Estimation Based on Deep Learning: A Survey 综述：基于深度学习的内窥镜深度估计技术，聚焦数据、方法与应用 depth estimation
16 Automated 3D-GS Registration and Fusion via Skeleton Alignment and Gaussian-Adaptive Features 提出基于骨架对齐和高斯自适应特征的3D-GS自动配准与融合方法 3D gaussian splatting gaussian splatting splatting
17 Reconstructing 4D Spatial Intelligence: A Survey 对4D空间智能重建进行综述，并按层级结构分析现有方法，展望未来方向。 scene reconstruction embodied AI ✅
18 Compositional Video Synthesis by Temporal Object-Centric Learning 提出基于时序对象中心学习的组合视频合成框架，实现高质量可控视频生成。 scene understanding

🔬 支柱一：机器人控制 (Robot Control) (4 篇)

#题目一句话要点标签🔗⭐
19 GaRe: Relightable 3D Gaussian Splatting for Outdoor Scenes from Unconstrained Photo Collections 提出基于3D高斯溅射的室外场景可重光照框架，解决无约束照片集的光照编辑问题。 manipulation 3D gaussian splatting gaussian splatting
20 FantasyID: A dataset for detecting digital manipulations of ID-documents FantasyID：用于检测身份证件数字篡改的公共数据集 manipulation
21 HAMLET-FFD: Hierarchical Adaptive Multi-modal Learning Embeddings Transformation for Face Forgery Detection 提出HAMLET-FFD，通过层级自适应多模态学习解决人脸伪造检测的跨域泛化问题。 manipulation
22 Learning to See Inside Opaque Liquid Containers using Speckle Vibrometry 提出基于散斑振动法的液位检测方法，用于识别不透明容器内的液体含量 manipulation

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

#题目一句话要点标签🔗⭐
23 Learning Only with Images: Visual Reinforcement Learning with Reasoning, Rendering, and Visual Feedback 提出RRVF框架，仅使用图像即可训练MLLM进行视觉推理，解决依赖图文标注问题 reinforcement learning large language model multimodal ✅
24 DAMS:Dual-Branch Adaptive Multiscale Spatiotemporal Framework for Video Anomaly Detection DAMS：用于视频异常检测的双分支自适应多尺度时空框架 visual pre-training spatiotemporal
25 Improving Adversarial Robustness Through Adaptive Learning-Driven Multi-Teacher Knowledge Distillation 提出自适应学习驱动的多教师知识蒸馏，提升CNN对抗鲁棒性 distillation
26 Beyond Class Tokens: LLM-guided Dominant Property Mining for Few-shot Classification 提出BCT-CLIP，利用LLM指导挖掘主导属性，提升少样本分类性能 representation learning contrastive learning

🔬 支柱七：动作重定向 (Motion Retargeting) (2 篇)

#题目一句话要点标签🔗⭐
27 Deep Learning for Skeleton Based Human Motion Rehabilitation Assessment: A Benchmark 提出Rehab-Pile康复数据集与基准框架，用于骨骼动作康复评估的深度学习研究 human motion
28 Top2Pano: Learning to Generate Indoor Panoramas from Top-Down View Top2Pano：提出一种从俯视图生成室内全景图的端到端模型 geometric consistency

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
29 MemoryTalker: Personalized Speech-Driven 3D Facial Animation via Audio-Guided Stylization MemoryTalker：通过音频引导的风格化实现个性化语音驱动的3D面部动画 motion synthesis

⬅️ 返回 cs.CV 首页 · 🏠 返回主页