Kaidong Zhang is a final year Ph.D candidate at University of Science and Technology of China (USTC). He is under the supervision of Dong Liu, and work closely with Jingjing Fu at MSRA. He also maintains a good cooperative relationship with the department of orthopedics, the first affiliated hospital of the university of science and technology of China.

His research interests include low-level vision, medical image processing and NeRF-based 3D rendering. He concentrated on video super-resolution and enhancement in the early period, and won the 2nd place in AI+4K HDR track of the first National Artificial Intelligence Competition (NAIC). Later, he transferred his research interest to video inpainting and editing, and published several papers on top conferences in computer vision. His paper FGT receives considerable attention from research community, and is reported by several computer vision forums, including VALSE and AI Time. His paper SAMed is a seminal work on transferring computer vision big models to medical image analysis area, and achieves more than 100 citations and 400 github stars. Now, he focuses more on the modeling of 3D reality in computer vision and the medical image processing for ground-breaking applications.

Kaidong Zhang received the B.Sc degree from Xi'an Jiaotong University in 2019. Here is the curriculum vitae.

News

• 2023.1.30 One paper is accepted by TPAMI 2024
• 2023.4.26 We release the paper and the codes of SAMed, which is the first implementation of Segment Anything Model (SAM) on medical image segmantic segmentation.
• 2023.4.24 FGT is reported by VALSE, one of the most influential Chinese forum for CV researchers.
• 2023.1.25 The journal extension of FGT (FGT++) is released on Arxiv.
• 2022.12.14 I was invited as a speaker to give a talk (in Chinese) about FGT in Youth PhD Talk ECCV-2022, hosted by AITime.
• 2022.10.14 I accomplished a talk (in Chinese) about our newly proposed video inpainting method "FGT" (accepted by ECCV 2022) in "AI新青年" lecture.
• 2022.7.4 One paper is accepted by ECCV 2022
• 2022.3.2 One paper is accepted by CVPR 2022