In October, LIU Ze from School of Information Science and Technology, University of Science and Technology of China (USTC) won the Best Paper Award, Marr Prize, at the 2021 International Conference on Computer Vision (ICCV). The conference paper was titled “Swin Transformer: Hierarchical Vision Transformer Using Shifted Windows”.
Swin Transformer presented in LIU Ze’s paper is the first Transformer-based model and capably serves as a general-purpose backbone for computer vision. Swin Transformer has refreshed the records significantly in object detection and semantic segmentation. Swin Transformer has been applied in a broad range of vision tasks in recent half year. The universality and effectiveness of Swin Transformer have been further proved in video action recognition, contrastive learning, image restoration and person re-identification.
LIU Ze, the lead author of this awarded paper, is currently a 3rd-year PhD student in USTC and Microsoft Research Asia (MSRA), under the co-supervision of Prof. GUO Baining and Prof. WANG Yong in the direction of deep learning and computer vision. He graduated with his bachelor degree from the Talent Program in the School of Information Science and Technology from USTC. LIU Ze has got the “GUO Moruo Scholarship”, the highest honor for students in USTC.
ICCV, the premier international computer vision event, was held from 11th to 17th of October, 2021. The high-profile Marr Prize of ICCV is one of the highest honors in computer vision and awarded every two years. This prize is named after the late British scientist David Marr, known as the “father of computer vision” and the founder of computational neuroscience. This year, ICCV received 1612 papers from 6152 submissions. Only LIU Ze’s paper was awarded the Marr Prize.
Article Title
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows