人工智能学院2021年系列学术活动——华为加拿大研究院温哥华研究所高级研究员马锐学术报告

Title: Intelligent and Controllable 3D Content Modeling

报告题目: 智能化和可控化的三维内容建模

Abstract: 3D models are the most essential elements in computer graphics and 3D vision applications such as 3D game, animation, VR and AR. However, creating and modeling the 3D content usually requires professional skills and is time-consuming. With the advances of machine learning and deep neural networks, data-driven 3D modeling becomes a promising solution to address the bottleneck for 3D content creation. In this talk, I will introduce my PhD and recent research on learning-based intelligent frameworks for creating novel 3D content in a controllable manner. Specifically, we focus on the 3D indoor scenes and man-made shapes which are challenging for modeling due to their complex geometry and structure. I will show how the proposed intelligent tools such as action-driven and natural language-driven scene synthesis as well as a deep structural-aware single-view reconstruction network can allow more accessible and more efficient 3D content modeling for novice users.

报告摘要：三维模型是计算机图形学和三维视觉相关应用（例如3D游戏、动画、VR和AR）中最必不可缺的元素。然而，三维内容的创建和建模通常需要专业的技巧并且耗时耗力。随着机器学习和深度神经网络的不断发展，数据驱动的三维建模逐渐成为解决三维内容创建瓶颈的有效方法。本报告将介绍我在博士期间和最近研究中提出的基于机器学习的智能化和可控化的三维内容创建方法。具体而言，我们将主要关注具有复杂几何和结构特征的三维室内场景和人造物体模型创建。我将展示如何利用基于动作和基于自然语言的场景生成，以及具有结构感知的单视图重建等智能工具，使新手用户也能快速上手并进行高效的三维建模。

报告人介绍：马锐，目前在华为加拿大研究院温哥华研究所担任高级研究员。2018 年博士毕业于加拿大西蒙菲莎大学计算机系，2013年和2010年在吉林大学数学学院分别获得计算数学硕士和信息与计算科学本科学位。主要研究方向包括计算机图形学，计算机视觉和机器学习等。在计算机图形学顶级会议Siggraph/Siggraph Asia上发表多篇论文。在公司工作期间发表并授权2项美国专利。曾担任IEEE Transactions on Visualization and Computer Graphics, Computer Graphics Forum, Graphical Models, AAAI, IEEE Virtual Reality, Graphics Interface 等图形学和AI领域国际期刊和会议的审稿人。

Bio: Rui Ma is a Senior Researcher in Huawei Vancouver Research Center, Canada. He obtained his PhD from School of Computing Science, Simon Fraser University in 2018, and MSc and BSc from School of Mathematics, Jilin University in 2013 and 2010, respectively. His main research interests include computer graphics, computer vision and machine learning. He has published several papers in Siggraph/Siggraph Asia, the most prestigious conference in computer graphics. He has also published and granted two US patents. Rui has served as a reviewer for graphics and AI journals and conferences such as IEEE Transactions on Visualization and Computer Graphics, Computer Graphics Forum, Graphical Models, AAAI, IEEE Virtual Reality, Graphics Interface.

报告时间：2021年5月24日（周一）上午9:10-10:10

报告平台：腾讯会议

会议号：197 876 528

会议密码:202104