cs.CV（2024-10-18）

📊 共 16 篇论文 | 🔗 4 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (8 🔗3) 支柱三：空间感知与语义 (Perception & Semantics) (3) 支柱二：RL算法与架构 (RL & Architecture) (3 🔗1) 支柱四：生成式动作 (Generative Motion) (1) 支柱六：视频提取与匹配 (Video Extraction) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (8 篇)

#题目一句话要点标签🔗⭐
1 Automating Video Thumbnails Selection and Generation with Multimodal and Multistage Analysis 提出一种多模态多阶段分析方法，自动选择和生成高质量视频缩略图。 large language model multimodal
2 Swiss Army Knife: Synergizing Biases in Knowledge from Vision Foundation Models for Multi-Task Learning 提出Swiss Army Knife，融合视觉基础模型知识偏见，提升多任务学习性能。 foundation model ✅
3 ViCToR: Improving Visual Comprehension via Token Reconstruction for Pretraining LMMs ViCToR：通过视觉Token重建提升LMMs的视觉理解能力 large language model multimodal ✅
4 Toward Generalizing Visual Brain Decoding to Unseen Subjects 提出一种通用的视觉脑解码框架，提升模型在未见个体上的泛化能力 foundation model ✅
5 Vision-Language Navigation with Energy-Based Policy 提出基于能量的导航策略以解决视觉语言导航问题 VLN
6 Storyboard guided Alignment for Fine-grained Video Action Recognition 提出基于故事板引导对齐的细粒度视频动作识别方法 large language model
7 Fine-Grained Verifiers: Preference Modeling as Next-token Prediction in Vision-Language Alignment 提出FiSAO，利用视觉编码器进行token级反馈，提升视觉-语言模型对齐效果 large language model
8 ProReason: Multi-Modal Proactive Reasoning with Decoupled Eyesight and Wisdom ProReason：解耦视觉感知与文本推理，实现多模态主动推理 large language model

🔬 支柱三：空间感知与语义 (Perception & Semantics) (3 篇)

#题目一句话要点标签🔗⭐
9 LUDVIG: Learning-Free Uplifting of 2D Visual Features to Gaussian Splatting Scenes LUDVIG：免学习地将2D视觉特征提升到高斯溅射场景 gaussian splatting splatting open-vocabulary
10 DaRePlane: Direction-aware Representations for Dynamic Scene Reconstruction DaRePlane：提出方向感知表示方法，用于动态场景重建，实现高保真新视角合成。 gaussian splatting splatting NeRF
11 Neural Signed Distance Function Inference through Splatting 3D Gaussians Pulled on Zero-Level Set 提出基于3D高斯溅射和神经SDF的表面重建方法，解决离散、稀疏和漂移问题。 3D gaussian splatting 3DGS gaussian splatting

🔬 支柱二：RL算法与架构 (RL & Architecture) (3 篇)

#题目一句话要点标签🔗⭐
12 Preview-based Category Contrastive Learning for Knowledge Distillation 提出基于预览的类别对比学习知识蒸馏方法，提升学生模型性能。 contrastive learning curriculum learning distillation
13 DRACO-DehazeNet: An Efficient Image Dehazing Network Combining Detail Recovery and a Novel Contrastive Learning Paradigm DRACO-DehazeNet：结合细节恢复和对比学习的高效图像去雾网络 contrastive learning ✅
14 MambaSCI: Efficient Mamba-UNet for Quad-Bayer Patterned Video Snapshot Compressive Imaging MambaSCI：用于Quad-Bayer视频快照压缩成像的高效Mamba-UNet Mamba

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
15 LEAD: Latent Realignment for Human Motion Diffusion 提出LEAD：通过潜在空间重对齐实现更真实的文本驱动人体运动扩散生成。 motion diffusion text-to-motion motion synthesis

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

#题目一句话要点标签🔗⭐
16 Multi-modal Pose Diffuser: A Multimodal Generative Conditional Pose Prior 提出MOPED：一种多模态条件扩散模型，作为SMPL姿态参数的先验，提升人体姿态生成质量。 SMPL multimodal

⬅️ 返回 cs.CV 首页 · 🏠 返回主页