cs.CV（2026-02-15）

📊 共 15 篇论文 | 🔗 3 篇有代码

🎯 兴趣领域导航

支柱二：RL算法与架构 (RL & Architecture) (5 🔗1) 支柱三：空间感知与语义 (Perception & Semantics) (4) 支柱九：具身大模型 (Embodied Foundation Models) (4 🔗2) 支柱七：动作重定向 (Motion Retargeting) (1) 支柱六：视频提取与匹配 (Video Extraction) (1)

🔬 支柱二：RL算法与架构 (RL & Architecture) (5 篇)

#题目一句话要点标签🔗⭐
1 UniWeTok: An Unified Binary Tokenizer with Codebook Size $\mathit{2^{128}}$ for Unified Multimodal Large Language Model 提出UniWeTok以解决多模态大语言模型的视觉表示问题 distillation large language model multimodal
2 LaViDa-R1: Advancing Reasoning for Unified Multimodal Diffusion Language Models 提出LaViDa-R1：一种用于统一多模态扩散语言模型的推理方法 reinforcement learning multimodal
3 ForgeryVCR: Visual-Centric Reasoning via Efficient Forensic Tools in MLLMs for Image Forgery Detection and Localization 提出ForgeryVCR，利用多模态大语言模型和取证工具实现图像伪造检测与定位 reinforcement learning large language model multimodal ✅
4 GeoEyes: On-Demand Visual Focusing for Evidence-Grounded Understanding of Ultra-High-Resolution Remote Sensing Imagery GeoEyes：针对超高分辨率遥感影像，提出按需视觉聚焦的证据驱动理解方法 reinforcement learning large language model multimodal
5 CoCoEdit: Content-Consistent Image Editing via Region Regularized Reinforcement Learning CoCoEdit：通过区域正则化强化学习实现内容一致的图像编辑 reinforcement learning

🔬 支柱三：空间感知与语义 (Perception & Semantics) (4 篇)

#题目一句话要点标签🔗⭐
6 Learnable Multi-level Discrete Wavelet Transforms for 3D Gaussian Splatting Frequency Modulation 提出可学习多层离散小波变换，用于3D高斯溅射的频率调制，减少高斯基元数量。 3D gaussian splatting 3DGS gaussian splatting
7 DenseMLLM: Standard Multimodal LLMs are Intrinsic Dense Predictors DenseMLLM：无需定制解码器的标准多模态LLM用于密集预测任务 depth estimation large language model multimodal
8 Flow4R: Unifying 4D Reconstruction and Tracking with Scene Flow Flow4R：提出统一的场景流框架，用于动态场景的4D重建与跟踪 scene understanding scene flow spatiotemporal
9 Dual-Signal Adaptive KV-Cache Optimization for Long-Form Video Understanding in Vision-Language Models Sali-Cache：面向长视频理解的视觉-语言模型双信号自适应KV-Cache优化 optical flow

🔬 支柱九：具身大模型 (Embodied Foundation Models) (4 篇)

#题目一句话要点标签🔗⭐
10 Towards Spatial Transcriptomics-driven Pathology Foundation Models 提出SEAL框架，通过空间转录组学信息增强病理学基础模型，提升视觉表征能力。 foundation model multimodal
11 BitDance: Scaling Autoregressive Generative Models with Binary Tokens BitDance：通过二元视觉令牌扩展自回归生成模型，实现高效图像生成。 foundation model multimodal ✅
12 MarsRetrieval: Benchmarking Vision-Language Models for Planetary-Scale Geospatial Retrieval on Mars MarsRetrieval：火星行星尺度地理空间检索的视觉-语言模型基准测试 foundation model multimodal ✅
13 HiVid: LLM-Guided Video Saliency For Content-Aware VOD And Live Streaming HiVid：利用LLM指导的视频显著性，实现内容感知的VOD和直播优化 large language model

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

#题目一句话要点标签🔗⭐
14 Differential pose optimization in descriptor space -- Combining Geometric and Photometric Methods for Motion Estimation 提出一种基于描述子空间的差分位姿优化方法，融合几何与光度信息进行运动估计 motion estimation

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

#题目一句话要点标签🔗⭐
15 EgoSound: Benchmarking Sound Understanding in Egocentric Videos EgoSound：首个用于评估以自我为中心视频中声音理解能力的基准 egocentric Ego4D large language model

⬅️ 返回 cs.CV 首页 · 🏠 返回主页