GAMES Webinar 2020 – 140期(计算机视觉专题) | Hao Su(UC San Diego)
GAMES Webinar 2020-140期】(计算机视觉专题)
报告嘉宾：Hao Su (UC San Diego)
报告题目：3D Understanding Towards Object Manipulation
3D understanding has been a critical component for autonomous systems that can interact with the environment. In the past few years, there have emerged a number of core algorithms that use deep learning to understand/create 3D data with unprecedented performance. It is, therefore, timely to ask 1) how these core algorithms may benefit downstream applications such as robotics; 2) what kind of new 3D deep learning problems may worth paying attention to; and 3) what kind of efforts may be needed to build new benchmarks. I will talk about some recent efforts in my group to explore the answers to the three questions. Specifically, I will talk about learning-based 3D capturing for robots (CVPR’20 oral), the 3D representation of antipodal grasps for object manipulation (CoRL’19), zero-shot object part discovery for novel categories (ICLR’20), and the creation of content-rich robot simulator (SAPIEN, CVPR’20 oral).
Hao Su has been in UC San Diego as Assistant Professor of Computer Science and Engineering since July 2017. He served as the Area Chair of ECCV’20, CVPR’20, ACCV’20, ICCV’19, CVPR’19, Senior Program Committee of AAAI’19, Program Chair of 3DV’17, and committee of other vision/graphics/learning conferences.
Professor Su is interested in fundamental problems in broad disciplines related to artificial intelligence, including machine learning, computer vision, computer graphics, and robotics. His representative work include ShapeNet, PointNet series, RenderForCNN, ObjectBank, ImageNet, etc. Recently, his interest is in theories and technologies to build autonomous systems that can actively and continuously learn in the physical world.
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；