3D 视觉 - arXivDaily 专题

1907.10844 2019-07-26 cs.CV 83%

PU-GAN: a Point Cloud Upsampling Adversarial Network

Ruihui Li, Xianzhi Li, Chi-Wing Fu, Daniel Cohen-Or, Pheng-Ann Heng

专题命中点云：point cloud(title,abstract)；3D reconstruction(abstract)；分类 cs.CV

Comments accepted by ICCV2019, project page at https://liruihui.github.io/publication/PU-GAN

1809.02743 2018-09-11 cs.CV 83%

RealPoint3D: Point Cloud Generation from a Single Image with Complex Background

Yan Xia, Yang Zhang, Dingfu Zhou, Xinyu Huang, Cheng Wang, Ruigang Yang

专题命中点云：point cloud(title,abstract)；3D generation(abstract)；分类 cs.CV

Comments 8 pages, 6 figures

2606.04376 2026-07-28 eess.IV cs.MM 版本更新 82%

FUSE-Flow: A Decoupled Framework for Calibration and Stateless Real-Time Multi-View Point Cloud Fusion

FUSE-Flow：一种用于标定和无状态实时多视角点云融合的解耦框架

Chentian Sun

专题命中点云：point cloud(title,abstract)；3D reconstruction(abstract)

AI总结提出一种解耦标定与融合的框架FUSE-Flow，通过几何对齐标定模块和置信度引导融合模块，实现无标定板、无全局优化的实时多视角点云融合，在视觉效果、动态稳定性和可扩展性上优于主流方法。

Comments 8pages,5figures, the version to submit IEEE TMM, resubmitted on 2026.7.8

2605.20978 2026-05-21 cs.LG 82%

Point Cloud Sequence Encoding for Material-conditioned Graph Network Simulators

用于材料条件化图网络模拟器的点云序列编码

Philipp Dahlinger, Balázs Gyenes, Niklas Freymuth, Luca Geminiani, Tobias Würth, Johannes Mitsch, Nadja Klein, Luise Kärger, Gerhard Neumann

机构 * Autonomous Learning Robots（自主学习机器人）； Methods for Big Data（大数据方法）； Institute of Vehicle System Technology（车辆系统技术研究所）

专题命中点云：point cloud(title,abstract)

AI总结本文提出PEACH框架，通过点云序列编码实现对未知物理属性的适应，提高了模拟到现实的零样本转移精度，并在实际部署中更具实用性。

Comments 9 pages + appendix, 7 figures. Submitted to the 40th Conference on Neural Information Processing Systems (NeurIPS 2026)

2404.05522 2025-01-10 cs.MM 82%

3DMambaIPF: A State Space Model for Iterative Point Cloud Filtering via Differentiable Rendering

Qingyuan Zhou, Weidong Yang, Ben Fei, Jingyi Xu, Rui Zhang, Keyi Liu, Yeqi Luo, Ying He

专题命中点云：point cloud(title,abstract)；3D vision(abstract)

Comments Accepted at AAAI-25

2412.02998 2024-12-05 cs.RO cs.CV cs.GR 82%

QuadricsReg: Large-Scale Point Cloud Registration using Quadric Primitives

Ji Wu, Huai Yu, Shu Han, Xi-Meng Cai, Ming-Feng Wang, Wen Yang, Gui-Song Xia

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.GR、cs.RO

Comments 25 pages, 17 figures

2410.16303 2024-10-23 eess.SP cs.LG 82%

Spatio-Temporal 3D Point Clouds from WiFi-CSI Data via Transformer Networks

Tuomas Määttä, Sasan Sharifipour, Miguel Bordallo López, Constantino Álvarez Casado

专题命中点云：point cloud(title,abstract)；3D reconstruction(abstract)

Comments 7 pages, 5 figures, 1 table

2308.05410 2023-08-11 cs.CV cs.GR cs.RO 82%

SC3K: Self-supervised and Coherent 3D Keypoints Estimation from Rotated, Noisy, and Decimated Point Cloud Data

Mohammad Zohaib, Alessio Del Bue

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.GR、cs.RO

Comments This paper has been accepted in International Conference on Computer Vision (ICCV) 2023. For code and data, please refer to the following GitHub page: https://github.com/IITPAVIS/SC3K

1912.12098 2020-08-25 cs.LG cs.CV cs.GR cs.RO stat.ML 82%

Quaternion Equivariant Capsule Networks for 3D Point Clouds

Yongheng Zhao, Tolga Birdal, Jan Eric Lenssen, Emanuele Menegatti, Leonidas Guibas, Federico Tombari

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.GR、cs.RO

Comments Oral Presentation at ECCV 2020. Find our video under: https://youtu.be/LHh56snwhTA. We release our sources at: http://tolgabirdal.github.io/qecnetworks

2006.09835 2020-06-18 eess.SP cs.IT cs.LG math.IT 82%

Wireless 3D Point Cloud Delivery Using Deep Graph Neural Networks

Takuya Fujihashi, Toshiaki Koike-Akino, Siheng Chen, Takashi Watanabe

专题命中点云：point cloud(title,abstract)；3D reconstruction(abstract)

Comments 5 pages

1912.11932 2019-12-30 cs.GR cs.CV cs.RO 82%

Skeleton Extraction from 3D Point Clouds by Decomposing the Object into Parts

Vijai Jayadevan, Edward Delp, Zygmunt Pizlo

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.GR、cs.RO

1811.07014 2019-11-05 cs.CV cs.GR cs.RO 82%

Topology-Aware Non-Rigid Point Cloud Registration

Konstantinos Zampogiannis, Cornelia Fermuller, Yiannis Aloimonos

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.GR、cs.RO

1906.03299 2019-10-01 cs.CV cs.GR cs.RO 82%

PyramNet: Point Cloud Pyramid Attention Network and Graph Embedding Module for Classification and Segmentation

Kang Zhiheng, Li Ning

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.GR、cs.RO

Comments Accepted for presentation at ICONIP2019

1909.03669 2019-09-10 cs.CV cs.AI cs.GR cs.RO 82%

DensePoint: Learning Densely Contextual Representation for Efficient Point Cloud Processing

Yongcheng Liu, Bin Fan, Gaofeng Meng, Jiwen Lu, Shiming Xiang, Chunhong Pan

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.GR、cs.RO

Comments Accepted to ICCV 2019. 15 pages, 8 figures, 16 tables

1903.03247 2019-03-11 cs.MM eess.SP 82%

HoloCast: Graph Signal Processing for Graceful Point Cloud Delivery

Takuya Fujihashi, Toshiaki Koike-Akino, Takashi Watanabe, Philip V. Orlik

专题命中点云：point cloud(title,abstract)；3D reconstruction(abstract)

1901.01255 2019-01-08 cs.CV cs.CG cs.GR cs.RO 82%

Generic Primitive Detection in Point Clouds Using Novel Minimal Quadric Fits

Tolga Birdal, Benjamin Busam, Nassir Navab, Slobodan Ilic, Peter Sturm

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.GR、cs.RO

Comments Submitted to IEEE Transactions on Pattern Analysis and Machine Intelligence (T-PAMI). arXiv admin note: substantial text overlap with arXiv:1803.07191

1807.00399 2018-11-19 cs.CV cs.CG cs.GR cs.RO 82%

cilantro: A Lean, Versatile, and Efficient Library for Point Cloud Data Processing

Konstantinos Zampogiannis, Cornelia Fermuller, Yiannis Aloimonos

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.GR、cs.RO

2106.11481 2021-06-23 cs.CV cs.AI cs.IR cs.LG cs.RO 82%

SeqNetVLAD vs PointNetVLAD: Image Sequence vs 3D Point Clouds for Day-Night Place Recognition

Sourav Garg, Michael Milford

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.RO；3D vision(comments)

Comments Accepted to CVPR 2021 Workshop on 3D Vision and Robotics (3DVR). https://sites.google.com/view/cvpr2021-3d-vision-robotics/

2011.00320 2020-11-03 cs.CV cs.LG cs.RO 82%

Scene Flow from Point Clouds with or without Learning

Jhony Kaesemodel Pontes, James Hays, Simon Lucey

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.RO；3D vision(comments)

Comments International Conference on 3D Vision (3DV 2020)

2512.03621 2026-08-05 cs.CV 版本更新 81%

ReCamDriving: LiDAR-Free Camera-Controlled Video Synthesis for Novel Trajectories

ReCamDriving：无需LiDAR的相机控制新型轨迹视频生成

Yaokun Li, Shuaixian Wang, Mantang Guo, Jiehui Huang, Taojun Ding, Mu Hu, Kaixuan Wang, Shaojie Shen, Guang Tan

机构 * Sun Yat-sen University（中山大学）； ZYT ； Shenzhen Polytechnic University（深圳职业技术大学）； The Hong Kong University of Science and Technology（香港科技大学）

专题命中点云：3DGS(summary_cn,abstract)；分类 cs.CV

AI总结 ReCamDriving通过基于3DGS的密集渲染和两阶段训练策略，实现了无需LiDAR的相机可控新型轨迹视频生成，构建了ParaDrive数据集并展示了卓越的生成效果。

Comments Project page: https://recamdriving.github.io/

2504.01732 2025-04-10 cs.CV 81%

FIORD: A Fisheye Indoor-Outdoor Dataset with LIDAR Ground Truth for 3D Scene Reconstruction and Benchmarking

Ulas Gunes, Matias Turkulainen, Xuqian Ren, Arno Solin, Juho Kannala, Esa Rahtu

专题命中点云：NeRF(abstract)；Gaussian Splatting(abstract)；point cloud(abstract)；novel view synthesis(abstract)

Comments SCIA 2025

2512.17897 2026-08-14 cs.CV cs.AI cs.LG cs.RO 版本更新 81%

RadarGen: Automotive Radar Point Cloud Generation from Cameras

RadarGen：从摄像头生成汽车雷达点云

Tomer Borreda, Fangqiang Ding, Sanja Fidler, Shengyu Huang, Or Litany

机构 * Technion（技术学院）； MIT（麻省理工学院）； NVIDIA（英伟达）； University of Toronto（多伦多大学）； Vector Institute（向量研究所）

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.RO

AI总结 RadarGen通过扩散模型从摄像头图像生成逼真的雷达点云，结合BEV对齐的深度、语义和运动线索，提升雷达生成的物理合理性与多模态模拟能力。

Comments ECCV 2026. Project page: https://radargen.github.io/

2602.11554 2026-08-11 cs.RO cs.CV cs.LG 版本更新 81%

HyperDet: 3D Object Detection with Hyper 4D Radar Point Clouds

HyperDet: 基于超4D雷达点云的3D目标检测

Yichun Xiao, Runwei Guan, Jin Jin, Fangqiang Ding

机构 * University of Edinburgh（爱丁堡大学）； HKUST (GZ)（香港科技大学（广州））； University of Oxford（牛津大学）； MIT（麻省理工学院）

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.RO

AI总结提出一种与检测器无关的框架HyperDet，通过构建任务感知的超4D雷达点云，利用时空累积、跨传感器验证和多普勒引导的运动补偿以及前景生成增强，显著提升仅用雷达的3D目标检测性能。

Comments 11 pages, 3 figures, 3 tables

2409.07558 2026-08-11 cs.CV cs.LG cs.RO 版本更新 81%

Unsupervised Point Cloud Registration with Self-Distillation

基于自蒸馏的无监督点云配准

Christian Löwens, Thorben Funke, André Wagner, Alexandru Paul Condurache

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.RO

AI总结本文针对点云配准依赖真实位姿标注的问题，提出自蒸馏无监督方法，在3DMatch基准上优于现有方法，且可泛化至汽车雷达场景。

Comments Oral at BMVC 2024

2607.28855 2026-08-03 cs.GR cs.CV 新提交 81%

Learning Manifolds in High-D Point Embedding for Anisotropic Surface Approximation from Unstructured Point Clouds

用于非结构化点云各向异性曲面逼近的高维点嵌入学习流形

Hongbo Li, Haikuan Zhu, Xiaohu Guo, Wenping Wang, Jing Hua, Zichun Zhong

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.GR

AI总结该研究提出HD-PEA框架，通过高维点嵌入等技术实现非结构化点云的各向异性曲面逼近，在多数据集上验证了其泛化性与可用性，性能优于现有方法。

2601.06442 2026-07-22 cs.CV cs.RO 版本更新 81%

WHU-PCPR: A cross-platform heterogeneous point cloud dataset for place recognition in complex urban scenes

WHU-PCPR：用于复杂城市场景中地点识别的跨平台异构点云数据集

Xianghong Zou, Jianping Li, Yandi Yang, Weitong Wu, Yuan Wang, Qiegen Liu, Zhen Dong

机构 * School of Advanced Manufacturing, Nanchang University（南昌大学先进制造学院）； School of Electrical and Electronic Engineering, Nanyang Technological University（南洋理工大学电子与电气工程学院）； Department of Geomatics Engineering, University of Calgary（卡尔加里大学测绘工程系）； School of Earth Sciences and Engineering, Hohai University（河海大学地球科学与工程学院）； School of Geography and Environment, Jiangxi Normal University（江西师范大学地理与环境学院）； School of Information Engineering, Nanchang University（南昌大学信息工程学院）； State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University（武汉大学测绘遥感信息工程国家重点实验室）； Hubei Luojia Laboratory（湖北省珞珈实验室）

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.RO

AI总结针对现有PCPR数据集在场景、平台和传感器方面缺乏多样性的问题，建立跨平台异构点云数据集WHU-PCPR，其具备独特特征。基于此对几种PCPR方法评估分析，讨论关键挑战与未来方向，推动相关研究发展。

2607.16946 2026-07-21 cs.GR cs.CV 新提交 81%

Points as Tori: Fast Pointwise Signed Distance for Point Clouds

点作为环面：用于点云的快速逐点符号距离

Nicole Feng, Ioannis Gkioulekas, Keenan Crane

机构 * Carnegie Mellon University（卡内基梅隆大学）

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.GR

AI总结该研究提出一种计算点云符号距离的方法，通过用环面局部拟合点云，利用预训练网络输出参数，无需全局优化和空间离散化且易并行化，基于新理论统一相关方法，可直接用于点云应用，无需显式重建。

Comments Published in ACM Transactions on Graphics, volume 45 (2026). For associated presentations and code, see https://nzfeng.github.io/research/PointsAsTori

Journal ref ACM Transactions on Graphics, volume 45 (2026), article number 53

2607.16146 2026-07-20 cs.RO cs.CV 新提交 81%

VTLoc: Learning-based Tactile Contact Localization in Visual Point Clouds

VTLoc：基于学习的视觉点云中触觉接触定位

Zhiyuan Wu, Zhuo Chen, Shan Luo

机构 * Department of Engineering, King’s College London（伦敦国王学院工程系）

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.RO

AI总结研究视觉与触觉融合的接触定位问题，提出VTLoc框架，通过几何多模态对齐模块和迭代定位更新器，利用视觉点云从触觉读数定位接触点，在新基准上减少对应模糊性，改善单触接触定位。

2607.14639 2026-07-17 cs.RO cs.CV 新提交 81%

Image-to-Point Cloud Registration Made Easy with Rectified Flow-based LiDAR Upsampling

基于整流流的激光雷达上采样实现图像到点云配准简化

Reon Tabata, Kenji Koide, Shuji Oishi, Masashi Yokozuka, Taku Okawara, Aoki Takanose, Jun Miura

机构 * Department of Computer Science and Engineering, Toyohashi University of Technology（丰桥技术科学大学计算机科学与工程系）； National Institute of Advanced Industrial Science and Technology（国立先进工业科学技术研究所）

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.RO

AI总结研究图像到点云配准难题，提出将激光雷达视为成像传感器的方法，通过条件整流流、特征匹配等步骤估计位姿，经自监督预训练和少量数据微调，实现高精度快速配准，性能优于现有方法。

2607.06782 2026-07-09 cs.RO cs.CV 新提交 81%

G-PROBE: Cross-FOV Place Recognition and Certainty-Coupled Localization for 3D Point Clouds

G-PROBE：用于3D点云的跨视场位置识别和确定性耦合定位

Jinseop Lee

机构 * SK Intellix（SK智能科技）

专题命中点云：point cloud(title,abstract)；分类 cs.CV、cs.RO

AI总结研究针对3D点云在有限或不对称视场下全局定位难的问题，提出G-PROBE框架。通过虚拟传感器分解、跨视场分支集合等方法，结合确定性耦合定位，无需学习。在多数据集和模式上评估，该框架性能出色，端到端可用性高，在不对称视场下优势明显。

Comments 18 pages, 9 figures

1. 点云 7702 篇

PU-GAN: a Point Cloud Upsampling Adversarial Network

RealPoint3D: Point Cloud Generation from a Single Image with Complex Background

FUSE-Flow: A Decoupled Framework for Calibration and Stateless Real-Time Multi-View Point Cloud Fusion

Point Cloud Sequence Encoding for Material-conditioned Graph Network Simulators

3DMambaIPF: A State Space Model for Iterative Point Cloud Filtering via Differentiable Rendering

QuadricsReg: Large-Scale Point Cloud Registration using Quadric Primitives

Spatio-Temporal 3D Point Clouds from WiFi-CSI Data via Transformer Networks

SC3K: Self-supervised and Coherent 3D Keypoints Estimation from Rotated, Noisy, and Decimated Point Cloud Data

Quaternion Equivariant Capsule Networks for 3D Point Clouds

Wireless 3D Point Cloud Delivery Using Deep Graph Neural Networks

Skeleton Extraction from 3D Point Clouds by Decomposing the Object into Parts

Topology-Aware Non-Rigid Point Cloud Registration

PyramNet: Point Cloud Pyramid Attention Network and Graph Embedding Module for Classification and Segmentation

DensePoint: Learning Densely Contextual Representation for Efficient Point Cloud Processing

HoloCast: Graph Signal Processing for Graceful Point Cloud Delivery

Generic Primitive Detection in Point Clouds Using Novel Minimal Quadric Fits

cilantro: A Lean, Versatile, and Efficient Library for Point Cloud Data Processing

SeqNetVLAD vs PointNetVLAD: Image Sequence vs 3D Point Clouds for Day-Night Place Recognition

Scene Flow from Point Clouds with or without Learning

ReCamDriving: LiDAR-Free Camera-Controlled Video Synthesis for Novel Trajectories

FIORD: A Fisheye Indoor-Outdoor Dataset with LIDAR Ground Truth for 3D Scene Reconstruction and Benchmarking

RadarGen: Automotive Radar Point Cloud Generation from Cameras

HyperDet: 3D Object Detection with Hyper 4D Radar Point Clouds

Unsupervised Point Cloud Registration with Self-Distillation

Learning Manifolds in High-D Point Embedding for Anisotropic Surface Approximation from Unstructured Point Clouds

WHU-PCPR: A cross-platform heterogeneous point cloud dataset for place recognition in complex urban scenes

Points as Tori: Fast Pointwise Signed Distance for Point Clouds

VTLoc: Learning-based Tactile Contact Localization in Visual Point Clouds

Image-to-Point Cloud Registration Made Easy with Rectified Flow-based LiDAR Upsampling

G-PROBE: Cross-FOV Place Recognition and Certainty-Coupled Localization for 3D Point Clouds