cs.CV（2024-03-29）

📊 共 34 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) (15 🔗4) 支柱九：具身大模型 (Embodied Foundation Models) (7 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (6 🔗2) 支柱七：动作重定向 (Motion Retargeting) (2) 支柱一：机器人控制 (Robot Control) (2 🔗2) 支柱四：生成式动作 (Generative Motion) (1) 支柱八：物理动画 (Physics-based Animation) (1 🔗1)

🔬 支柱三：空间感知与语义 (Perception & Semantics) (15 篇)

#题目一句话要点标签🔗⭐
1 Snap-it, Tap-it, Splat-it: Tactile-Informed 3D Gaussian Splatting for Reconstructing Challenging Surfaces 提出触觉信息3D高斯点云重建方法以解决复杂表面重建问题 3D gaussian splatting 3DGS gaussian splatting
2 SGD: Street View Synthesis with Gaussian Splatting and Diffusion Prior 提出基于扩散模型的3D高斯点云合成方法以提升街景新视角渲染质量 3D gaussian splatting 3DGS gaussian splatting
3 HO-Gaussian: Hybrid Optimization of 3D Gaussian Splatting for Urban Scenes 提出HO-Gaussian以解决城市场景3D高斯点云渲染问题 3D gaussian splatting 3DGS gaussian splatting
4 InstantSplat: Sparse-view Gaussian Splatting in Seconds 提出InstantSplat以解决稀疏视图3D重建问题 3D reconstruction gaussian splatting splatting
5 FreeSeg-Diff: Training-Free Open-Vocabulary Segmentation with Diffusion Models 提出FreeSeg-Diff以解决开放词汇图像分割问题 open-vocabulary open vocabulary foundation model ✅
6 HGS-Mapping: Online Dense Mapping Using Hybrid Gaussian Representation in Urban Scenes 提出HGS-Mapping以解决城市场景在线稠密映射问题 3D gaussian splatting 3DGS gaussian splatting
7 NeSLAM: Neural Implicit Mapping and Self-Supervised Feature Tracking With Depth Completion and Denoising 提出NeSLAM以解决RGB-D传感器深度图稀疏与噪声问题 depth estimation 3D reconstruction NeRF
8 SCINeRF: Neural Radiance Fields from a Snapshot Compressive Image 提出SCINeRF以从快照压缩图像中恢复3D场景表示 NeRF neural radiance field ✅
9 MI-NeRF: Learning a Single Face NeRF from Multiple Identities 提出MI-NeRF以解决多身份动态人脸建模问题 NeRF neural radiance field
10 VSRD: Instance-Aware Volumetric Silhouette Rendering for Weakly Supervised 3D Object Detection 提出VSRD框架以解决弱监督3D目标检测问题 depth estimation monocular depth scene understanding ✅
11 Stable Surface Regularization for Fast Few-Shot NeRF 提出稳定表面正则化方法以加速少样本NeRF合成 NeRF
12 SceneTracker: Long-term Scene Flow Estimation Network 提出SceneTracker以解决长期场景流估计问题 scene flow ✅
13 Talk3D: High-Fidelity Talking Portrait Synthesis via Personalized 3D Generative Prior 提出Talk3D以解决单目视频面部几何重建不足问题 NeRF neural radiance field
14 DerainNeRF: 3D Scene Estimation with Adhesive Waterdrop Removal 提出DerainNeRF以解决雨雪天气下图像水滴干扰问题 NeRF neural radiance field
15 Sparse Views, Near Light: A Practical Paradigm for Uncalibrated Point-light Photometric Stereo 提出无校准多视角光度立体方法以解决稀疏视角重建问题 3D reconstruction

🔬 支柱九：具身大模型 (Embodied Foundation Models) (7 篇)

#题目一句话要点标签🔗⭐
16 Unsolvable Problem Detection: Robust Understanding Evaluation for Large Multimodal Models 提出UPD以评估大型多模态模型的理解能力 multimodal chain-of-thought ✅
17 Mixed-precision Supernet Training from Vision Foundation Models using Low Rank Adapter 提出混合精度超网训练以优化视觉基础模型的部署 foundation model
18 Uncovering Bias in Large Vision-Language Models with Counterfactuals 通过反事实分析揭示大型视觉语言模型中的偏见 large language model multimodal
19 Draw-and-Understand: Leveraging Visual Prompts to Enable MLLMs to Comprehend What You Want 提出Draw-and-Understand框架以增强多模态大语言模型的理解能力 large language model multimodal
20 Convolutional Prompting meets Language Models for Continual Learning 提出ConvPrompt以解决持续学习中的知识共享问题 large language model
21 Learn "No" to Say "Yes" Better: Improving Vision-Language Models via Negations 提出CC-Neg数据集与CoN-CLIP框架以改善视觉语言模型对否定的理解 foundation model
22 Long-Tailed Anomaly Detection with Learnable Class Names 提出LTAD方法以解决长尾异常检测问题 foundation model

🔬 支柱二：RL算法与架构 (RL & Architecture) (6 篇)

#题目一句话要点标签🔗⭐
23 Unsupervised Tumor-Aware Distillation for Multi-Modal Brain Image Translation 提出UTAD-Net以解决多模态脑影像翻译中的肿瘤变形问题 teacher-student distillation ✅
24 UltraLight VM-UNet: Parallel Vision Mamba Significantly Reduces Parameters for Skin Lesion Segmentation 提出UltraLight VM-UNet以解决皮肤病变分割中的参数冗余问题 Mamba SSM ✅
25 HARMamba: Efficient and Lightweight Wearable Sensor Human Activity Recognition Based on Bidirectional Mamba 提出HARMamba以解决可穿戴传感器人类活动识别的效率问题 Mamba
26 Colorful Cutout: Enhancing Image Data Augmentation with Curriculum Learning 提出Colorful Cutout以增强图像数据增强效果 curriculum learning
27 Heterogeneous Network Based Contrastive Learning Method for PolSAR Land Cover Classification 提出异构网络对比学习方法以解决PolSAR分类问题 contrastive learning
28 Learning Enriched Features via Selective State Spaces Model for Efficient Image Deblurring 提出选择性状态空间模型以解决高效图像去模糊问题 SSM state space model

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

#题目一句话要点标签🔗⭐
29 Motion Inversion for Video Customization 提出运动嵌入以解决视频生成中的运动定制问题 motion representation spatiotemporal
30 A Unified Framework for Human-centric Point Cloud Video Understanding 提出统一框架以解决人类中心点云视频理解问题 human motion

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
31 ECLIPSE: Efficient Continual Learning in Panoptic Segmentation with Visual Prompt Tuning 提出ECLIPSE以解决持续学习中的灾难性遗忘问题 manipulation distillation ✅
32 P-Hologen: An End-to-End Generative Framework for Phase-Only Holograms 提出P-Hologen以解决全光相干全息图生成问题 manipulation ✅

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
33 CT respiratory motion synthesis using joint supervised and adversarial learning 提出一种深度合成方法以生成伪呼吸CT相位 motion synthesis

🔬 支柱八：物理动画 (Physics-based Animation) (1 篇)

#题目一句话要点标签🔗⭐
34 Video-Based Human Pose Regression via Decoupled Space-Time Aggregation 提出DSTA网络以解决视频中人类姿态估计的计算瓶颈问题 spatiotemporal ✅

⬅️ 返回 cs.CV 首页 · 🏠 返回主页