cs.CV（2025-09-05）

📊 共 20 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

支柱三：空间感知与语义 (Perception & Semantics) (7 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (5 🔗1) 支柱九：具身大模型 (Embodied Foundation Models) (4) 支柱一：机器人控制 (Robot Control) (2) 支柱五：交互与反应 (Interaction & Reaction) (1) 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) (7 篇)

#题目一句话要点标签🔗⭐
1 Visibility-Aware Language Aggregation for Open-Vocabulary Segmentation in 3D Gaussian Splatting 提出VALA，解决3D高斯溅射开放词汇分割中背景噪声和多视角不一致问题 3D gaussian splatting gaussian splatting splatting
2 FlowSeek: Optical Flow Made Easier with Depth Foundation Models and Motion Bases FlowSeek：利用深度基础模型和运动基的光流估计框架，降低训练成本并提升泛化性 optical flow foundation model
3 CoRe-GS: Coarse-to-Refined Gaussian Splatting with Semantic Object Focus CoRe-GS：面向语义兴趣点的粗到精高斯溅射，加速移动重建。 gaussian splatting splatting scene reconstruction
4 GeoSplat: A Deep Dive into Geometry-Constrained Gaussian Splatting GeoSplat：提出几何约束高斯溅射框架，提升新视角合成性能 gaussian splatting splatting
5 FloodVision: Urban Flood Depth Estimation Using Foundation Vision-Language Models and Domain Knowledge Graph FloodVision：结合视觉语言模型与领域知识图谱的城市洪水深度估计 depth estimation
6 SGS-3D: High-Fidelity 3D Instance Segmentation via Reliable Semantic Mask Splitting and Growing SGS-3D：通过可靠语义掩码分割与生长实现高保真3D实例分割 scene understanding
7 A biologically inspired separable learning vision model for real-time traffic object perception in Dark 提出一种生物启发式可分离学习视觉模型，用于黑暗环境下的实时交通目标感知。 optical flow ✅

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

#题目一句话要点标签🔗⭐
8 Symbolic Graphics Programming with Large Language Models 提出基于强化学习的框架，提升大语言模型生成精确可控SVG图像的能力 reinforcement learning large language model
9 PropVG: End-to-End Proposal-Driven Visual Grounding with Multi-Granularity Discrimination PropVG：提出端到端的基于提议的视觉定位框架，提升复杂场景下的目标识别能力。 contrastive learning visual grounding ✅
10 DuoCLR: Dual-Surrogate Contrastive Learning for Skeleton-based Human Action Segmentation DuoCLR：通过双代理对比学习增强骨骼动作分割 representation learning contrastive learning
11 SL-SLR: Self-Supervised Representation Learning for Sign Language Recognition 提出SL-SLR框架，通过自监督学习提升手语识别的表征能力 representation learning contrastive learning
12 Dynamic Sensitivity Filter Pruning using Multi-Agent Reinforcement Learning For DCNN's 提出差分敏感度融合剪枝算法，用于高效压缩深度卷积神经网络 reinforcement learning

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

#题目一句话要点标签🔗⭐
13 Efficient Video-to-Audio Generation via Multiple Foundation Models Mapper 提出多基础模型映射器(MFM-Mapper)，高效生成与视频内容匹配的音频。 foundation model
14 UniView: Enhancing Novel View Synthesis From A Single Image By Unifying Reference Features UniView：通过统一参考特征增强单图像的新视角合成 large language model multimodal
15 MCANet: A Multi-Scale Class-Specific Attention Network for Multi-Label Post-Hurricane Damage Assessment using UAV Imagery 提出MCANet，利用多尺度类特定注意力网络进行无人机图像的飓风灾后多标签评估。 large language model multimodal
16 WatchHAR: Real-time On-device Human Activity Recognition System for Smartwatches WatchHAR：面向智能手表的实时、端侧人体活动识别系统 multimodal

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
17 OpenEgo: A Large-Scale Multimodal Egocentric Dataset for Dexterous Manipulation OpenEgo：用于灵巧操作的大规模多模态第一人称数据集 manipulation dexterous hand dexterous manipulation
18 LUIVITON: Learned Universal Interoperable VIrtual Try-ON LUIVITON：学习的通用互操作虚拟试穿系统，适用于复杂服装和多样化人体 humanoid SMPL foundation model

🔬 支柱五：交互与反应 (Interaction & Reaction) (1 篇)

#题目一句话要点标签🔗⭐
19 Scale-interaction transformer: a hybrid cnn-transformer model for facial beauty prediction 提出Scale-Interaction Transformer (SIT)模型，用于提升面部美学预测的准确性。 interaction transformer

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

#题目一句话要点标签🔗⭐
20 Comparative Evaluation of Traditional and Deep Learning Feature Matching Algorithms using Chandrayaan-2 Lunar Data 利用嫦娥二号月球数据，对比传统与深度学习特征匹配算法，实现精准图像配准。 feature matching

⬅️ 返回 cs.CV 首页 · 🏠 返回主页