点云三维重建的最新进展

Ying He 教授

现任新加坡南洋理工大学计算机科学与数据科学学院副教授，主要研究方向为几何计算与分析，并在计算机图形学与视觉领域有广泛应用。他担任《IEEE可视化与计算机图形学汇刊》《计算机图形学论坛》及《计算视觉媒体》等期刊编委，长期活跃于几何建模领域主要国际会议的技术程序委员会，并多次担任重要会议主席或程序委员会联合主席，包括：国际形状建模会议（2022年）、实体与物理建模研讨会（2022年、2023年）、几何建模与处理会议（2014年、2021年）、计算视觉媒体会议（2020年）及太平洋图形学会议（2026年）。

从点云重建三维曲面是数字几何处理与三维视觉领域的基础性问题。在众多方法中，基于隐函数的方法因其能恢复复杂拓扑结构且对噪声具有鲁棒性而广受关注，其中泊松曲面重建及其变体仍是最广泛应用的技术之一。传统方法通常假设输入点云具有全局一致的方向性，这使得点云定向成为密切相关且关键的子问题。长期以来，该领域的研究主要沿两个独立方向推进：一是聚焦于估计一致的法向方向，二是基于定向点云进行曲面重建。本次报告将介绍一系列近期研究项目，这些工作将两个方向统一为协调一致的框架，实现对方向估计与曲面重建的联合求解。最后，我将探讨这一快速发展领域的新兴趋势与开放挑战。

王智慧教授

2025年12月19日（星期五）

上午10:00-12:00

深圳河套学院B411阶梯教室

（深圳市福田区福保街道红棉路6号，

地图导航“深圳河套学院-南门”）

扫码加入会议

腾讯会议号：758-956-403

SLAI

Recent Advances in 3D Reconstruction from Point Clouds

Ying He

Ying He is an Associate Professor in the College of Computing and Data Science at Nanyang Technological University, Singapore. His research interests focus on geometric computing and analysis, with applications in computer graphics and vision. He serves on the editorial boards of IEEE Transactions on Visualization and Computer Graphics, Computer Graphics Forum, and Computational Visual Media, and has been an active member of the technical program committees for major conferences in geometric modeling. He has also served, or will serve, as General or Program Co-Chair for several conferences, including Shape Modeling International (2022), the Symposium on Solid and Physical Modeling (2022, 2023), Geometric Modeling and Processing (2014, 2021), Computational Visual Media (2020), and Pacific Graphics (2026).

Reconstructing 3D surfaces from point clouds is a fundamental problem in digital geometry processing and 3D vision. Among the various approaches, implicit function–based methods have become particularly popular for their ability to recover complex topologies and their robustness against noise. Poisson surface reconstruction and its variants remain among the most widely used techniques. Traditionally, these methods assume that input point clouds are equipped with globally consistent orientations, making point orientation a closely related and critical subproblem. Until recently, research in this area largely advanced along two separate directions: one focusing on estimating consistent normal orientations, and the other on reconstructing surfaces from oriented points. In this talk, I will present a series of recent projects that unify these two threads into a single, coherent framework that jointly addresses orientation estimation and surface reconstruction. I will conclude with a discussion of emerging trends and open challenges in this rapidly evolving field.

Prof. Wang Zhihui

December 19, 2025, Friday,

10:00am-12:00pm

B411 Lecture Hall,Shenzhen Loop Area Institute

(6 Hongmian Rd, Fubao Sub-Street, Futian District, Shenzhen, navigate to "Shenzhen Loop Area Institute (South Gate)" on the map.)

Join the Meeting Online

Tencent Meeting：758-956-403