11/12(五)_Computer vision: past, present and future_主講人:Harry Shum, Ph.D.
國立清華大學資訊工程學系
Department of Computer Science
National Tsing Hua University
專題演講
SEMINAR
主講人： Harry Shum, Ph.D.
SPEAKER (Adjunct Professor, Tsinghua University
Foreign Member, National Academy of Engineering, U.S.A.
International Fellow, Royal Academy of Engineering, U.K.)
題 目：Computer vision: past, present and future
TOPIC
時 間：110年11月12日(五)上午10點10分至12點
地 點：台達館103
Abstract
The ubiquitous presence of digital cameras in our daily lives and ever-increasing computing power have enabled many applications for computer vision from taking panoramic images on mobile phones today to autonomous driving in the future. In this talk, I will start with a brief overview of the history of computer vision. Since “artificial intelligence” was coined in 1956, computer vision researchers have been hard at work pursuing numerous representations of images for various vision tasks, from 3D reconstruction to object detection and recognition. Notable examples include Marr’s 2.5D sketch, Canny edge detector, convolutional neural networks (CNN), SIFT Keypoint detector, etc. In the last decade, deep CNN’s have become the most popular representation, exemplified by ResNet, pretrained on large image data sets like ImageNet. Recently, vision transformers have captured a lot of imaginations by leveraging more data and relaxing convolution constraints of CNN. Indeed, unified vision and language representation learning will be key to image understanding tasks from image captioning and multi-modal dialog. The future of computer vision will continue to rely on our relentless search for better representations of images and tasks.
Bio:
Dr. Harry Shum is an adjunct professor at Institute for Advanced Study, Tsinghua University. He was executive vice president of Microsoft’s Artificial Intelligence (AI) and Research group. He was responsible for driving the company’s overall AI strategy and forward-looking research and development efforts spanning infrastructure, services, apps and agents. He oversaw AI-focused product groups including Bing. He also led Microsoft Research, one of the world’s premier computer science research organizations, and its integration with the engineering teams across the company.
Host: Yu-Ting Kuo