Short Bio:
Feng Xu
Professor, Vice Dean
Key Lab for Information Science of Electromagnetic Waves (MoE)
School of Information Science and Technology
Fudan University
Shanghai, China
Prof. Xu received the B.E. with honor in Information Engineering from Southeast University, Nanjing, China and the Ph.D. with honor in Electronic Engineering from Fudan University, Shanghai, China, in 2003 and 2008, respectively. From 2008 to 2010, he was a postdoctoral fellow with the NOAA Center for Satellite Application and Research (STAR), Camp Springs, MD. From 2010 to 2013, he was with at Intelligent Automation Inc. Rockville MD, while partly working for NASA Goddard Space Flight Center, Greenbelt, MD as a research scientist. In 2012, he was selected into China’s Global Experts Recruitment Program, and subsequently returned to Fudan University in June 2013, where he currently is a professor in the school of information science and technology and the vice director of the MoE Key Lab for Information Science of Electromagnetic Waves. He has published more than 60 papers in peer-reviewed journals and co-authored 3 books, among many conference papers. Among other honors, he was awarded the second-class National Nature Science Award of China in 2011. He was the 2014 recipient of the Early Career Award of IEEE Geoscience and Remote Sensing Society and the 2007 recipient of the SUMMA graduate fellowship in the advanced electromagnetics area. He currently serves as the associate editor for IEEE Geoscience and Remote Sensing Letters. He is the founding chair of IEEE GRSS Shanghai Chapter and member of IEEE GRSS AdCom. His research interests include electromagnetic scattering modeling, SAR information retrieval and radar system development.

Title: Deep Learning Methods and Applications in SAR Image Interpretation

Abstract: In the big data era of earth observation, deep learning and other data mining technologies become critical to successful end applications. Deep learning technology has revolutionized the computer vision areas, and is gradually being applied in radar remote sensing. Over the past several years, there has been exponentially increasing interests related to deep learning techniques applied to synthetic aperture radar (SAR) imagery. However, there are issues that are specific to SAR image interpretation such as limited training samples, sensitivity to observation configuration, or weak generalization ability. There are some techniques that can be used to mitigate these issues such as fusing electromagnetic physics laws with deep neural networks, using prior constraints of physical laws to realize few-shot learning capability, etc. This talk reports the recent progresses of the author and collaborators in this area.