At the 2019 IEEE CVPR，the Intelligent Voice team led by Associate Professor DU Jun from the University of Science and Technology of China（USTC）, won a championship and a runner-up in the object detection competition under optical remote sensing image.
The team from USTC proposed an adjustable period coding method to better represent the periodicity of the angle, and a stacked area convolution network based on length-independent overlap, making the model more robust to objects of different proportional lengths. In the competition, they won the 1st place in the task-object detection with oriented bounding boxes and won the 2nd place in the task-object detection with horizontal bounding boxes. This is another outstanding achievement after the team won the Objects Detecting in Aerial Images (ODAI) at the 2018 International Conference on Pattern Recognition (ICPR)，demonstrating the team's international research preponderance.
Image1，2: The Award Certificates of USTC at CVPR
In recent years, with the rapid development of artificial intelligence, object detection has become an important basic task of computer vision. Object detection under optical remote sensing image has important application value for land planning, disaster monitoring and other fields of science.
The CVPR’s purpose is to encourage in-depth discussion of these technical and application issues. In the past few years, the workshop proposals have proved so popular an d numerous that conference organizers have only been able to allow only 90. This object detection competition of remote sensing-image was jointly organized by Wuhan University, Huazhong University of Science and Technology, Cornell University, Rochester University, German Aerospace Center and other institutions, attracting many Chinese universities and institutions to participate it, including USTC, Nanjing University of Science and Technology, Institute of Electronics, Chinese Academy of Sciences, University of Chinese Academy of Sciences, Dalian University of Technology, Harbin Institute of Technology, Xi'an University of Electronic Science and Technology etc.
National Engineering Laboratory for Speech and Language Information Processing(NEL-SLIP)
(Written by ZHANG Xin,edited by LI Xiaoxi, USTC News Center)