GAMES Webinar 2021 – 205期(几何专题) | Bo Yang (The Hong Kong Polytechnic University)

【GAMES Webinar 2021-205期】(几何专题,Talk+Panel形式)

报告嘉宾:Bo Yang (The Hong Kong Polytechnic University)

报告时间:2021年10月28号星期四晚上8:00-9:00(北京时间)

报告题目:Learning 3D Representations without Full Supervision

报告摘要:

A long-standing goal in machine intelligence is to build systems that are able to infer the underlying 3D scene structure as well as understand its semantic composition. This is a fundamental necessity for autonomous driving, augmented reality, and robotics. Core problems include shape reconstruction, semantic segmentation, and instance segmentation for large-scale 3D scenes. At the beginning of this talk, I will briefly summarise our previous fully-supervised methods for 3D shape reconstruction and 3D scene understanding. After that, I will present our recent works using weakly-supervised and unsupervised methods for 3D representation learning. At last, the key challenges and possible research directions will be discussed.

讲者简介:

Dr. Bo Yang is an Assistant Professor and leads the Visual Learning and Reasoning Group in the Department of Computing at The Hong Kong Polytechnic University. He completed his DPhil(PhD) degree in computer science at the University of Oxford. He is generally interested in machine learning, computer vision, and robotics. The overarching goal of his research is to build intelligent systems which endow machines to recover, understand, and eventually interact with the real 3D world.

讲者个人主页:https://yang7879.github.io/


主持人简介:

徐凯,国防科技大学教授。2011年于国防科大计算机学院获得博士学位。普林斯顿大学访问学者。西蒙弗雷泽大学客座教授。研究方向为数据驱动的三维感知与建模、三维视觉及其机器人应用等。发表CCF A类论文50余篇,其中包括领域顶级会议和期刊ACM SIGGRAPH/Transactions on Graphics论文20余篇。担任ACM Transactions on Graphics、Computer Graphics Forum、Computers and Graphics和The Visual Computer等期刊的编委。担任CAD/Graphics 2017、ISVC 2018等国际会议的论文共同主席,以及SIGGRAPH、SIGGRAPH Asia、Eurographics等会议的程序委员。担任中国工业与应用数学学会理事、中国图学学会理事、中国图象图形学会三维视觉专委会副主任、中国工业与应用数学学会几何设计与计算专委会秘书长。获湖南省自然科学一等奖、中国计算机学会自然科学一等奖、军队科技进步二等奖等。获国家优秀青年科学基金和湖南省杰出青年科学基金资助。


Panelists简介:

智帅峰是伦敦帝国理工学院-戴森机器人实验室的四年级博士研究生,导师为Andrew Davison教授和Stefan Leutenegger教授,本科和硕士毕业于国防科技大学。其研究方向和博士课题主要关注面向场景理解和语义SLAM系统的语义场景表征学习,探索和构筑适合Spatial AI系统所需的语义-几何信息的联合表征。

郭裕兰,国防科技大学副教授。主要研究方向为三维点云特征学习、三维场景感知与理解。目前已在IEEE TPAMI、IJCV和CVPR等国际期刊和会议上发表学术论文100余篇,论文被引用5000余次。目前担任中国图象图形学学会三维视觉专委会秘书长, CVPR 2021、ICCV 2021和ACM MM 2021等国际会议领域主席,中国图象图形学报,Computer & Graphics等期刊编委,曾担任IEEE TPAMI期刊客座编辑。曾获得Elsevier中国高被引学者,ACM China SIGAI Rising Star Award,中国人工智能学会吴文俊人工智能优秀青年奖,中国人工智能学会优秀博士学位论文。曾入选中国首届博士后创新人才支持计划。

 

GAMES主页的“使用教程”中有 “如何观看GAMES Webinar直播?”及“如何加入GAMES微信群?”的信息;
GAMES主页的“资源分享”有往届的直播讲座的视频及PPT等。
观看直播的链接:http://webinar.games-cn.org

You may also like...