<rt id="oaiu8"><small id="oaiu8"></small></rt>
<acronym id="oaiu8"><center id="oaiu8"></center></acronym>
當前位置: 首頁 >> 培養信息 >> 正文

【講座預告】Attention-based Domain Adaptation for Hyperspectral Image Classification

發布時間:2020-12-13  發布作者:  點擊數:

網絡在線講座:Attention-based Domain Adaptation for Hyperspectral Image Classification

主講人:Qian (Jenny) Du,Department of Electrical and Computer Engineering, Mississippi State University, USA




主講人簡介:Dr. Du is Bobby Shackouls Professor with the Department of Electrical and Computer Engineering, Mississippi State University, USA. Her research interests include hyperspectral remote sensing image analysis and applications, pattern recognition, and machine learning. Dr. Du is a Fellow of IEEE and SPIE—International Society for Optics and Photonics. She served as Co-Chair for the Data Fusion Technical Committee of IEEE Geoscience and Remote Sensing Society (GRSS) in 2009–2013, and Chair for Remote Sensing and Mapping Technical Committee of International Association for Pattern Recognition (IAPR) in 2010–2014. She served as Associate Editor for IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (2011–2015), IEEE Signal Processing Letters (2012–2015), and Journal of Applied Remote Sensing (2014– 2015). Dr. Du is the Editor-in-Chief of IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (JSTARS) in 2016-2020. She was the Guest Editor of several special issues published in remote sensing journals. Dr. Du is the General Chair of the 4th IEEE GRSS Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS), the 7th and 8th IAPR Workshop on Pattern Recognition in Remote Sensing (PRRS), and the 2019 IEEE Workshop on the Analysis of Multitemporal Remote Sensing Images (MultiTemp). She is also the Technical Co-chair of the 2019 International Conference on Agro-Geoinformatics. She is a member of IEEE TAB Periodicals Review and Advisory Committee (PRAC) and SPIE Publications Committee.

講座簡介:Machine learning algorithms have been extensively used to extract complex features for classification of remotely sensed hyperspectral images. However, for challenging tasks, such as domain adaptation, traditional algorithms tend to perform less effectively. Recently, with the advent of deep learning algorithms, more complex but useful features can be generated for hyperspectral image classification, and such algorithms often have a certain level of domain adaptation capability. However, attention-based feature generation is not fully explored till now, which has been found to be effective for distinguishing classes in different images than without transferring such information. In this talk, attention-based transfer learning is discussed for domain adaptation of hyperspectral imagery, where different levels and types of attention are transferred from a supervisor network (i.e., deep residual network) to a student network (i.e., wide residual network) to provide important features for improving the overall classification of the wide residual network. It is expected that the trained wide residual network is simpler but more efficient, and it can offer excellent domain adaptation capability. It has been shown that the proposed attention-based transfer learning technique outperforms the state-of-the-art domain adaptation methods.