报告人： Jungong Han（Aberystwyth University）

报告题目：Open World Deep Learning for Visual Understanding

报告时间：2021年12月9日 19:00-20:30

报告地点：腾讯会议：934-169-672

报告摘要：In the last decade, research for visual understanding has become more prevalent due to the great success of deep learning, especially deep convolutional neural networks (DCNN). By feeding high-quality annotated training data into a fully supervised learning (FSL) engine, DCNN models could even surpass human-level performance in many visual understanding tasks, such as object classification and face recognition. However, conducting FSL in real-world scenarios is challenging due to 1) deep learning technique generates high-dimensional visual features, which make applications like real-time feature matching and large-scale retrieval intractable; 2) there are potentially unlimited object categories in real life such that it is almost impossible to collect enough well-annotated samples for each category; 3) existing DCNN solutions often require a large number of computational resources to run, which are not available on real-life embedded devices. In this talk, I will share with you three PhD projects that I supervised in the past 3 years, where we showcase how we tackled these three problems.

报告人简介：Jungong Han is a Chair Professor and the Director of Research of Computer Science at Aberystwyth University, UK. He also holds an Honorary Professorship at the University of Warwick. Han’s research interests span various topics of computer vision and video analytics, including object detection, tracking and recognition, human behavior analysis, and video semantic analysis. With his research students, he has published over 70 IEEE/ACM Transactions papers, and 19 conference papers from CVPR/ICCV/ECCV, NeurIPS, and ICML. His work has been well-received earning over 8400 citations and his H-index is 47 in Google scholar. He is the Associate Editor-in-Chief of Elsevier Neurocomputing, and the Associate Editor of Elsevier Pattern Recognition.