3D 视觉 - arXivDaily 专题

2411.18898 2024-12-02 cs.CV 80%

Textured As-Is BIM via GIS-informed Point Cloud Segmentation

Mohamed S. H. Alabassy

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

Comments Permission granted by all co-authors for the publication of the extended article to the conference paper "BIM Integration for Automated Identification of Relevant Geo-Context Information via Point Cloud Segmentation" (2023). URL: https://www.ucl.ac.uk/bartlett/construction/sites/bartlett_construction/files/bim_integration_for_automated_identification_of_relevant_geo-context.pdf

2411.03725 2024-11-07 cs.CV 80%

PX2Tooth: Reconstructing the 3D Point Cloud Teeth from a Single Panoramic X-ray

Wen Ma, Huikai Wu, Zikai Xiao, Yang Feng, Jian Wu, Zuozhu Liu

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

Comments Ma W, Wu H, Xiao Z, et al. PX2Tooth: Reconstructing the 3D Point Cloud Teeth from a Single Panoramic X-Ray[C]//International Conference on Medical Image Computing and Computer-Assisted Intervention. Cham: Springer Nature Switzerland, 2024: 411-421

2407.05862 2024-07-09 cs.CV 80%

Bringing Masked Autoencoders Explicit Contrastive Properties for Point Cloud Self-Supervised Learning

Bin Ren, Guofeng Mei, Danda Pani Paudel, Weijie Wang, Yawei Li, Mengyuan Liu, Rita Cucchiara, Luc Van Gool, Nicu Sebe

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

Comments Bringing Masked Autoencoders Explicit Contrastive Properties for Point Cloud Self-Supervised Learning

2306.16020 2023-06-29 cs.CV 80%

Points for Energy Renovation (PointER): A LiDAR-Derived Point Cloud Dataset of One Million English Buildings Linked to Energy Characteristics

Sebastian Krapf, Kevin Mayer, Martin Fischer

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

Comments The PointER dataset can be downloaded from https://doi.org/10.14459/2023mp1713501. The code used for generating building point clouds is available at https://github.com/kdmayer/PointER

2301.11630 2023-01-30 cs.CV eess.IV eess.SP 80%

Joint Geometry and Attribute Upsampling of Point Clouds Using Frequency-Selective Models with Overlapped Support

Viktoria Heimann, Andreas Spruck, André Kaup

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

Comments 10 pages, 10 figures, Under Review at IEEE TMM Special Issue on Point Cloud Processing and Understanding

2209.04401 2022-09-12 eess.IV cs.CV 80%

GRASP-Net: Geometric Residual Analysis and Synthesis for Point Cloud Compression

Jiahao Pang, Muhammad Asad Lodhi, Dong Tian

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

Comments Accepted at ACM MM 2022 Workshop on Advances in Point Cloud Compression, Processing and Analysis

2204.09337 2022-04-22 cs.CV 80%

Sequential Point Clouds: A Survey

Haiyan Wang, Yingli Tian

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

Comments Survey paper on sequential point clouds

2110.09129 2021-10-19 cs.CV 80%

Deep Models with Fusion Strategies for MVP Point Cloud Registration

Lifa Zhu, Changwei Lin, Dongrui Liu, Xin Li, Francisco Gómez-Fernández

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

Comments Point cloud registration competition, ICCV21 workshop. Substantial text overlap with arXiv:2107.02583

2108.00620 2021-10-15 cs.CV 80%

Investigating Attention Mechanism in 3D Point Cloud Object Detection

Shi Qiu, Yunfan Wu, Saeed Anwar, Chongyi Li

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

Comments International Conference on 3D Vision (3DV 2021)

2010.08719 2021-08-27 cs.CV cs.AI 80%

Cascaded Refinement Network for Point Cloud Completion with Self-supervision

Xiaogang Wang, Marcelo H Ang, Gim Hee Lee

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

Comments Accepted by PAMI. Extended version of the following paper: Cascaded Refinement Network for Point Cloud Completion. CVPR 2020. arXiv link: arXiv:2004.03327

1902.05247 2021-03-24 cs.CV 80%

3D Graph Embedding Learning with a Structure-aware Loss Function for Point Cloud Semantic Instance Segmentation

Zhidong Liang, Ming Yang, Chunxiang Wang

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

Journal ref Title: 3D Instance Embedding Learning With a Structure-Aware Loss Function for Point Cloud Segmentation; Published in: IEEE Robotics and Automation Letters ( Volume: 5, Issue: 3, July 2020)

2101.00483 2021-01-05 cs.CV 80%

Learning Rotation-Invariant Representations of Point Clouds Using Aligned Edge Convolutional Neural Networks

Junming Zhang, Ming-Yuan Yu, Ram Vasudevan, Matthew Johnson-Roberson

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

Comments 3D Vision Conference 2020

2010.05391 2020-10-13 cs.CV cs.AI cs.LG 80%

A Progressive Conditional Generative Adversarial Network for Generating Dense and Colored 3D Point Clouds

Mohammad Samiul Arshad, William J. Beksi

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

Comments To be published in the 2020 International Conference on 3D Vision (3DV)

1908.08854 2020-06-30 cs.CV cs.LG eess.IV 80%

Linking Points With Labels in 3D: A Review of Point Cloud Semantic Segmentation

Yuxing Xie, Jiaojiao Tian, Xiao Xiang Zhu

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

Comments The title of published version was modified to "Linking Points With Labels in 3D: A Review of Point Cloud Semantic Segmentation". To read its final version please go to IEEE Geoscience and Remote Sensing Magazine on IEEE XPlore: https://ieeexplore.ieee.org/document/9028090

2005.11626 2020-05-26 cs.CV cs.LG eess.IV 80%

ShapeAdv: Generating Shape-Aware Adversarial 3D Point Clouds

Kibok Lee, Zhuoyuan Chen, Xinchen Yan, Raquel Urtasun, Ersin Yumer

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

Comments 3D Point Clouds, Adversarial Learning

1908.06297 2019-08-20 cs.CV 80%

Rotation Invariant Convolutions for 3D Point Clouds Deep Learning

Zhiyuan Zhang, Binh-Son Hua, David W. Rosen, Sai-Kit Yeung

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

Comments International Conference on 3D Vision (3DV) 2019

1812.09084 2018-12-24 cs.RO 80%

Casualty Detection from 3D Point Cloud Data for Autonomous Ground Mobile Rescue Robots

Roni Permana Saputra, Petar Kormushev

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

Comments Published in SSRR 2018 Conference

Journal ref R. P. Saputra and P. Kormushev, "Casualty Detection from 3D Point Cloud Data for Autonomous Ground Mobile Rescue Robots," 2018 IEEE International Symposium on Safety, Security, and Rescue Robotics (SSRR), Philadelphia, PA, 2018, pp. 1-7

1808.02350 2018-08-08 cs.CV eess.IV 80%

YOLO3D: End-to-end real-time 3D Oriented Object Bounding Box Detection from LiDAR Point Cloud

Waleed Ali, Sherif Abdelkarim, Mohamed Zahran, Mahmoud Zidan, Ahmad El Sallab

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

Comments Paper accepted in ECCV 2018, "3D Reconstruction meets Semantics" workshop

1608.05143 2017-06-07 cs.CV 80%

A Systematic Approach for Cross-source Point Cloud Registration by Preserving Macro and Micro Structures

Xiaoshui Huang, Jian Zhang, Lixin Fan, Qiang Wu, Chun Yuan

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

Comments Cross-source point cloud registration

2608.18627 2026-08-20 cs.CV 新提交 79%

PCQA-R1: Advancing Generalized 3D Point Cloud Quality Assessment with Reinforcement Learning

PCQA-R1：基于强化学习的通用三维点云质量评估方法

Kangning Ye, Yunhao Li, Sijing Wu, Yucheng Zhu, Guangtao Zhai

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

AI总结本文提出首个用于三维点云质量评估的强化学习LMM PCQA-R1，基于GRPO策略构建PCQA-CoT数据集并引入高斯 proximity 奖励，在跨数据集泛化和域内准确率上表现优异。

2608.18479 2026-08-20 cs.CV 新提交 79%

COSTA: A Cluster-Centric Paradigm for Annotation-Free Open-Set Semantic Segmentation of Aerial Point Clouds with Domain Shifts

COSTA：面向存在域偏移的航拍点云无标注开放集语义分割的以聚类为中心范式

Yanghong Lin, Li Fang, Tianyu Li, Shudong Zhou, Wei Yao

机构 * Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences（中国科学院福建物质结构研究所）； Institute of Urban Environment, Chinese Academy of Sciences（中国科学院城市环境研究所）； University of Chinese Academy of Sciences（中国科学院大学）； School of Resource and Environmental Sciences, Wuhan University（武汉大学资源与环境科学学院）

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

AI总结 COSTA提出以聚类为中心的范式，无需额外训练即可在推理阶段将预训练航拍点云分割模型适配到偏移目标域，在三个基准上实现按需分割，mIoU达70.09%。

2606.01549 2026-08-20 cs.CV 版本更新 79%

ForestMamba: Sparse Mamba with Geometry-guided Queries for 3D Forest Point Cloud Segmentation

ForestMamba: 基于几何引导查询的稀疏Mamba用于3D森林点云分割

Trung Thanh Nguyen, Tuan-Anh Vu, Duc Viet Le, Yasutomo Kawanishi, Takahiro Komamizu, Ichiro Ide, Teja Kattenborn

机构 * Nagoya University（名古屋大学）； RIKEN Seika（日本理化学研究所Seika研究中心）； University of California, Los Angeles（加州大学洛杉矶分校）； University of Twente（埃因霍温理工大学）； Ritsumeikan University（立命馆大学）

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

AI总结提出ForestMamba方法，通过稀疏编码器、几何引导查询初始化和Mamba查询解码器，实现高效且结构感知的森林点云分割，在七个森林区域上优于现有方法，推理速度提升3倍，GPU内存降低2.3倍。

Comments 37th British Machine Vision Conference (BMVC 2026)

2504.08154 2026-08-20 cs.CV 79%

Investigating Vision-Language Model for Point Cloud-based Vehicle Classification

Yiqiao Li, Jie Wei, Camille Kamga

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

Comments 5 pages,3 figures, 1 table, CVPR DriveX workshop

Journal ref DriveX Workshop at IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026

2608.08115 2026-08-19 cs.CV 版本更新 79%

SUMI: Scalable Unified Model for 3D Point Cloud Inference

SUMI：用于三维点云推理的可扩展统一模型

Yanlong Li, Kanchana Thilakarathna

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

AI总结针对点云补全由粗到精范式中局部细节重构不足的问题，提出扩散增强细化模块SUMI，可集成到现有模型，在多个基准数据集上实现性能提升。

2608.16225 2026-08-18 cs.CV 新提交 79%

PCT-Prompt: A Prompt-Guided Transformer Framework for Dense Prediction Tasks in Point Clouds

PCT-Prompt：用于点云密集预测任务的提示引导Transformer框架

Dejun Zhang, Yanzi Bai, Yiqi Wu

机构 * China University of Geosciences（中国地质大学）； Li Auto Inc（理想汽车）

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

AI总结 PCT-Prompt是一种提示引导Transformer框架，通过引入提示引导特征分支和提示丢弃机制，提升了标准Transformer在点云密集预测任务中的适应性，在多个公开数据集上取得了优异性能。

2608.15104 2026-08-18 cs.CV 新提交 79%

ProjFormer: Point Cloud Completion via Geometric-Projective Transformer and Cross-Modal Semantic Constraints

ProjFormer：基于几何投影Transformer与跨模态语义约束的点云补全

Sheng Liu, Meng Wang, Ruihui Li, Huilong Pi, Zhuo Tang, Kenli Li

机构 * Hunan University（湖南大学）

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

AI总结针对现有点云补全方法几何一致性与适应性不足的问题，提出ProjFormer跨模态框架，通过投影引导视图注意力模块与几何感知路由网络实现高效特征聚合与融合，在轻量级设计下取得了具竞争力的补全性能。

Comments Accepted by ACM Multimedia 2026. 10 pages, 6 figures, 5 tables

2608.14394 2026-08-17 cs.CV 新提交 79%

IRGNN: Efficient Invariant Radar Graph Neural Network for Radar Point Cloud Object Detection

IRGNN：用于雷达点云目标检测的高效不变雷达图神经网络

Xiao Guo, Wanke Xia, Lili Yang, Caicong Wu

机构 * China Agricultural University（中国农业大学）； Tsinghua University（清华大学）

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

AI总结针对雷达点云稀疏无序、信息不足难以适配现有激光雷达方法的问题，提出IRGNN，通过不变特征设计、改进MPNN及任务头实现目标检测，在RadarScenes数据集上性能更优且推理效率更高。

Comments Accepted at ICONIP 2026

2608.11697 2026-08-13 cs.CV 新提交 79%

Boundary-Enhanced Segmentation of Pig Point Clouds in Commercial Housing Environments

商业养殖环境下猪只点云的边界增强分割

Zhankang Xu, Fei Shi, Xiangyu Qi, Zhaoyang Wang, Mengxin Guo, Yikai Fan, Simon X. Yang, Qifeng Li, Weihong Ma

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

AI总结针对商业猪舍中猪只点云分割的边界模糊等问题，采用Octree Transformer骨干网络，结合软距离边界伪标签与双向跨边界语义模块，提升分割性能，为精准畜牧养殖提供可靠输入。

Comments 24 pages,9 figures, 5 tables

2608.09400 2026-08-11 cs.CV cs.AI 新提交 79%

Sign Language Recognition Using Original and Synthetic Depth Image Based Point Cloud Data Models

基于原始和合成深度图像的点云数据模型的手语识别

Rustem Ozakar, Eyup Gedikli

机构 * Erzurum Technical University（埃尔祖鲁姆技术大学）； Trabzon University（特拉布宗大学）

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

AI总结该研究利用Depth Anything V2生成合成深度图像，结合三个手语数据集，采用PointNet等模型对比原始与合成深度图像点云的手语识别性能，发现多数模型中原始数据性能更优，部分模型合成数据更优。

2608.08955 2026-08-11 cs.CV 新提交 79%

Damage Classification for 3D Point Cloud Data via 3D Data Analysis and Vision Foundation Model-based 2D Projections

基于三维数据分析与视觉基础模型二维投影的三维点云数据损伤分类

Evan Perez, Kalelo Dukuray, Erika Ardiles-Cruz, Jie Wei

机构 * City College of New York（纽约城市学院）； Air Force Research Lab（空军研究实验室）

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

AI总结本研究针对三维点云损伤分类的挑战，提出3PDA与2PDA两种算法，2PDA准确率略低但计算成本大幅降低且泛化性更强，为相关任务提供了高效替代方案。

1. 点云 7719 篇

Textured As-Is BIM via GIS-informed Point Cloud Segmentation

PX2Tooth: Reconstructing the 3D Point Cloud Teeth from a Single Panoramic X-ray

Bringing Masked Autoencoders Explicit Contrastive Properties for Point Cloud Self-Supervised Learning

Points for Energy Renovation (PointER): A LiDAR-Derived Point Cloud Dataset of One Million English Buildings Linked to Energy Characteristics

Joint Geometry and Attribute Upsampling of Point Clouds Using Frequency-Selective Models with Overlapped Support

GRASP-Net: Geometric Residual Analysis and Synthesis for Point Cloud Compression

Sequential Point Clouds: A Survey

Deep Models with Fusion Strategies for MVP Point Cloud Registration

Investigating Attention Mechanism in 3D Point Cloud Object Detection

Cascaded Refinement Network for Point Cloud Completion with Self-supervision

3D Graph Embedding Learning with a Structure-aware Loss Function for Point Cloud Semantic Instance Segmentation

Learning Rotation-Invariant Representations of Point Clouds Using Aligned Edge Convolutional Neural Networks

A Progressive Conditional Generative Adversarial Network for Generating Dense and Colored 3D Point Clouds

Linking Points With Labels in 3D: A Review of Point Cloud Semantic Segmentation

ShapeAdv: Generating Shape-Aware Adversarial 3D Point Clouds

Rotation Invariant Convolutions for 3D Point Clouds Deep Learning

Casualty Detection from 3D Point Cloud Data for Autonomous Ground Mobile Rescue Robots

YOLO3D: End-to-end real-time 3D Oriented Object Bounding Box Detection from LiDAR Point Cloud

A Systematic Approach for Cross-source Point Cloud Registration by Preserving Macro and Micro Structures

PCQA-R1: Advancing Generalized 3D Point Cloud Quality Assessment with Reinforcement Learning

COSTA: A Cluster-Centric Paradigm for Annotation-Free Open-Set Semantic Segmentation of Aerial Point Clouds with Domain Shifts

ForestMamba: Sparse Mamba with Geometry-guided Queries for 3D Forest Point Cloud Segmentation

Investigating Vision-Language Model for Point Cloud-based Vehicle Classification

SUMI: Scalable Unified Model for 3D Point Cloud Inference

PCT-Prompt: A Prompt-Guided Transformer Framework for Dense Prediction Tasks in Point Clouds

ProjFormer: Point Cloud Completion via Geometric-Projective Transformer and Cross-Modal Semantic Constraints

IRGNN: Efficient Invariant Radar Graph Neural Network for Radar Point Cloud Object Detection

Boundary-Enhanced Segmentation of Pig Point Clouds in Commercial Housing Environments

Sign Language Recognition Using Original and Synthetic Depth Image Based Point Cloud Data Models

Damage Classification for 3D Point Cloud Data via 3D Data Analysis and Vision Foundation Model-based 2D Projections