I am currently a Research Fellow at the College of Computing and Data Science (CCDS), Nanyang Technological University (NTU), working with Prof. Shijian Lu. I obtained my Ph.D., M.S., and B.S. degrees from Beihang University in 2025, 2022, and 2019, respectively, under the supervision of Prof. Zhenwei Shi and Prof. Zhengxia Zou.
Research & Impact: My research focuses on Computer Vision and Remote Sensing, with a particular emphasis on Foundation Models, Multimodal Learning, and AI4Earth. I have published over 40 papers in top-tier venues, including Proc. IEEE, TPAMI, GRSM, TGRS, and CVPR. My work has garnered over 9,000 citations (Google Scholar) and features 6 ESI Hot Papers and 13 ESI Highly Cited Papers. Notably, 4 of my works ranked among the Top 3 “Most Popular Papers” in their respective journals.
Academic Service: I have delivered invited talks at IEEE GRSS, the VALSE Student Forum (8 selected nationwide), and the CSIG Student Forum. My research has been featured by authoritative media such as IEEE and CSIG, and my open-source projects have earned 4,000+ GitHub stars. I serve as a Guest Editor for Remote Sensing and regularly review for ACM CS, IJCV, CVPR, TGRS, and TIP.
Leadership & Honors: I served as the Principal Investigator for the first batch of the NSFC Ph.D. Student Project. My algorithms have been successfully deployed in the Gaofen series and commercial satellite systems. I am the recipient of the Grand Prize of Baosteel Education Award (Top 10 graduate students nationwide), the Shen Yuan Medal (highest honor at Beihang), and the May Fourth Medal (highest youth honor at Beihang). Furthermore, I have been recognized by elite technical talent programs, including Alibaba (Ali-Star), Tencent (Qingyun), and JD.com (TGT).
I have authored over 50 peer-reviewed papers, featuring 6 ESI Hot Papers (Top 0.1%), 13 ESI Highly Cited Papers (Top 1%), and 5 Most Popular Articles (Top 10). Cumulative citations: 9432 on Google Scholar. Please refer to the Google Scholar profile or Full Publications List for all papers.
My research has received recognition from over 40 Academicians and Fellows, including a Turing Award winner. To date, my work has accumulated 9,000+ citations on Google Scholar and has been featured by authoritative organizations such as IEEE and CSIG. My open-source projects have garnered 4,000+ Stars on GitHub. Representative endorsements include:
Cited by Turing Award Winner: Prof. Yoshua Bengio (Turing Award Winner) cited my work multiple times in a NeurIPS’25 paper, noting that his research was “building upon insights from” my work and adopting it as a key research foundation to achieve significant improvements in his domain.
Recognized by Meta FAIR: The Meta FAIR Team (including Computer Vision pioneer/Distinguished Scientist Ross Girshick, Citations: 580k+) highlighted my work as one of only two representative methods in Remote Sensing within their SAM 2 technical report.
Included in Stanford CS231n Curriculum: My survey on object detection, a Most Popular Article in Proceedings of the IEEE, was selected as reference material for the world-renowned course “CS231n” (2023-2025) taught by Prof. Li Fei-Fei.
Praised by IEEE/AAIA Fellow: Prof. Jocelyn Chanussot (Director at INRIA, France) highly commended the RSPrompter method, stating it “successfully” and “autonomously” adapts large models for instance segmentation tasks, fully affirming the algorithm’s automation capabilities.
Recommended by IEEE/ACM/AAAS Fellow: Academician Dacheng Tao and Prof. Xudong Jiang (IEEE Fellow) introduced my work as a representative method in the top-tier journal TPAMI (2024). They assessed that the method “is capable of simultaneously detecting objects and their attributes”, validating the complementarity of fine-grained attributes for open-world detection.
Affirmed by Fellow of RSC/CAE/EIC: Academician Jonathan Li pointed out in his paper (JAG, 2023) that my algorithm designed a “compact” dual-channel architecture to achieve “efficiency maximization”, successfully establishing spatiotemporal long-range dependencies to enhance feature extraction.
Highlighted by Domain Experts: Prof. Fahad Shahbaz Khan dedicated a separate section to introduce my work in a Remote Sensing survey. Prof. Liangpei Zhang (IEEE Fellow) evaluated the adaptive prompting method in a NeurIPS paper as being “significantly different from other manual prompting methods”, possessing unique advantages.