cs.CV（2026-06-26）

📊 共 37 篇论文 | 🔗 10 篇有代码

🎯 兴趣领域导航

支柱三：空间感知与语义 (Perception & Semantics) (9 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (8 🔗2) 支柱九：具身大模型 (Embodied Foundation Models) (6 🔗2) 支柱七：动作重定向 (Motion Retargeting) (4 🔗1) 支柱一：机器人控制 (Robot Control) (4) 支柱四：生成式动作 (Generative Motion) (3 🔗2) 支柱八：物理动画 (Physics-based Animation) (2) 支柱六：视频提取与匹配 (Video Extraction) (1)

🔬 支柱三：空间感知与语义 (Perception & Semantics) (9 篇)

#题目一句话要点标签🔗⭐
1 StructSplat: Generalizable 3D Gaussian Splatting from Uncalibrated Sparse Views 提出StructSplat以解决无标定稀疏视图下的3D重建问题 3D gaussian splatting gaussian splatting splatting ✅
2 Understanding How MLLMs Describe Artworks Using Token Activation Maps 提出Token Activation Map以解析多模态大语言模型对艺术作品的描述 open-vocabulary open vocabulary large language model ✅
3 Panoramic Scene Analysis: A Survey from Distortion-Aware Engineering to Sphere-Native Foundation Modeling 提出全景场景分析方法以解决几何失真问题 depth estimation foundation model ✅
4 Every Step of the Way: Video-based Parkinsonian Turning Step Counting 提出视频基础的帕金森转身步数计数方法以解决现有技术不足 optical flow human mesh recovery motion representation
5 ReScene: Structured Indoor Scene Reconstruction from Multi-View Captures 提出ReScene框架以解决多视角室内场景重建问题 scene reconstruction physically plausible
6 GeoFace: Consistent Multi-View Face Generation with Geometry-Constrained Diffusion 提出GeoFace以解决多视角人脸生成中的几何一致性问题 3D reconstruction geometric consistency
7 Scene and Human in One World: Reconstruction in a Feedforward Pass 提出SHOW框架以解决动态场景中人类重建问题 scene reconstruction human mesh recovery
8 STAG: Spatio-temporal Evolving Structural Representation of Action Units for Micro-expression Recognition 提出STAG以解决微表情识别中的时空信息建模问题 optical flow
9 ScaLe-INR: Scale and Learn Implicit Neural Representations 提出ScaLe-INR以解决隐式神经表示中的频谱偏差问题 3D reconstruction

🔬 支柱二：RL算法与架构 (RL & Architecture) (8 篇)

#题目一句话要点标签🔗⭐
10 Directing the World: Fast Autoregressive Video Generation with Compositional Human-Camera Control 提出快速自回归框架以解决视频生成中的控制问题 world model world models human motion ✅
11 RPM-Distill: Physiology-guided Adaptive Cross-modal Distillation for Robust Remote Physiological Measurement 提出RPM-Distill以解决视频生理测量在复杂环境下的脆弱性问题 MAE distillation ✅
12 TempAct: Advancing Temporal Plausibility in Autoregressive Video Generation via Planner-Executor RL 提出TempAct框架以解决自回归视频生成中的时间一致性问题 reinforcement learning distillation instruction following
13 Exposure Bias Can Alleviate Itself via Directional and Frequency Rectification in Flow Matching 提出DEFAR框架以自我修正流匹配中的曝光偏差问题 flow matching
14 Reflect-R1: Evidence-Driven Reflection for Self-Correction in Long Video Understanding 提出Reflect-R1以解决长视频理解中的自我修正问题 reinforcement learning multimodal
15 ZooClaw-FashionSigLIP2: Distilled Fine-tuning for Robust Fashion Retrieval 提出ZooClaw-FashionSigLIP2以解决时尚检索中的模型泛化问题 distillation foundation model
16 Curriculum-guided Change Detection Training: Toward Accurate Serac Fall Monitoring 提出课程引导的变化检测训练以解决监测精度问题 representation learning curriculum learning
17 MASS: Motion-Aligned Selective Scan for Refinement in Flow-Based Video Frame Interpolation 提出MASS以解决视频帧插值中的运动对齐问题 SSM state space model

🔬 支柱九：具身大模型 (Embodied Foundation Models) (6 篇)

#题目一句话要点标签🔗⭐
18 RSICCLLM: A Multimodal Large Language Model for Remote Sensing Image Change Captioning 提出RSICCLLM以解决遥感图像变化描述问题 large language model multimodal ✅
19 PerceptionRubrics: Calibrating Multimodal Evaluation to Human Perception 提出PerceptionRubrics以解决多模态评估与人类感知之间的差距问题 multimodal
20 ProMSA:Progressive Multimodal Search Agents for Knowledge-Based Visual Question Answering 提出ProMSA以解决知识驱动视觉问答中的检索适应性问题 multimodal ✅
21 Video-MME-Logical: A Controlled Diagnostic Benchmark for Video Temporal-Logical Reasoning 提出Video-MME-Logical以解决视频时序逻辑推理问题 large language model multimodal
22 TextDS: Parameter-Efficient Representation Alignment for Scene Text Detection under Distribution Shifts 提出TextDS以解决场景文本检测中的分布转移问题 foundation model
23 Temporal-Emerged Prompting for Segment Anything in Multiframe Infrared Small Target Detection 提出TEP-SAM以解决低信噪比红外序列中的小目标检测问题 foundation model

🔬 支柱七：动作重定向 (Motion Retargeting) (4 篇)

#题目一句话要点标签🔗⭐
24 Long-Term Prediction of Local and Global Human Motion with Occlusion Recovery 提出非自回归变换器以解决人类运动预测中的局部与全局问题 human motion motion prediction ✅
25 AirGroundBench: Probing Spatial Intelligence in Multimodal Large Models under Heterogeneous Multi-View Embodied Collaboration 提出AirGroundBench以评估多模态大模型在异构多视角协作中的空间智能 geometric consistency large language model multimodal
26 HumanMoveVQA: Can Video MLLMs reason about human movement in videos? 提出HumanMoveVQA以解决视频中人类运动理解问题 human motion large language model multimodal
27 Learning Topology-Aware Representations via Test-Time Adaptation for Anomaly Segmentation 提出TopoTTA以解决异常分割中的结构一致性问题 spatial relationship

🔬 支柱一：机器人控制 (Robot Control) (4 篇)

#题目一句话要点标签🔗⭐
28 CascadeOcc: Rethinking 3D Occupancy World Models with Cascaded VQ Representations 提出CascadeOcc以解决3D占用世界模型的结构性不足问题 motion planning world model world models
29 ModaFlow: Modality-Aware Flow Matching for High-Fidelity Virtual Try-On 提出ModaFlow以解决高保真虚拟试穿中的多模态匹配问题 manipulation flow matching classifier-free guidance
30 Cross-view Multimodal Vision-Based Assessment Framework for Traditional Chinese Medicine Rehabilitation Training 提出CME-AQA框架以解决中医康复训练中的动作质量评估问题 manipulation multimodal
31 PhysisForcing: Physics Reinforced World Simulator for Robotic Manipulation 提出PhysisForcing以解决机器人操作中的物理一致性问题 manipulation world model world models

🔬 支柱四：生成式动作 (Generative Motion) (3 篇)

#题目一句话要点标签🔗⭐
32 HAT-4D: Lifting Monocular Video for 4D Multi-Object Interactions via Human-Agent Collaboration 提出HAT-4D以解决单目视频中多物体交互重建问题 physically plausible embodied AI VLA ✅
33 SIFT: Self-Imagination Fine-Tuning for Physically Plausible Motion in Video Diffusion Models 提出自我想象微调方法以解决视频扩散模型中的物理合理性问题 physically plausible
34 OrthoTryOn: Geometric Orthogonalization for Conflict-Free Unified Fashion Generation 提出OrthoTryOn以解决统一时尚生成中的任务冲突问题 classifier-free guidance ✅

🔬 支柱八：物理动画 (Physics-based Animation) (2 篇)

#题目一句话要点标签🔗⭐
35 TRUST: Efficient Abdominal Trauma Recognition via Image-to-Ultrasound-Video Transfer Learning 提出TRUST框架以解决腹部创伤超声视频理解问题 spatiotemporal
36 Enhancing Co-packaging Optics Enabled Silicon Photonics Security Assurance Hardware Fingerprinting 提出硬件指纹技术以解决光子集成电路安全问题 PHC

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

#题目一句话要点标签🔗⭐
37 EMOSH: Expressive Motion and Shape Disentanglement for Human Animation 提出EMOSH以解决人类动画中的运动与形状纠缠问题 SMPL

⬅️ 返回 cs.CV 首页 · 🏠 返回主页