Chun Yuan

Tsinghua University, Peng Cheng Laboratory

Papers

Total Citations

H-Index

About

Chun Yuan is a leading researcher in 3D computer vision and visual place recognition, with a focus on point cloud registration and hierarchical localization systems. His work addresses critical challenges in aligning cross-source 3D data—point clouds captured by different sensors—where variations in resolution, noise, and viewpoint make registration difficult. His 2016 paper on a coarse-to-fine algorithm for 3D street-view point cloud registration (28 citations) laid foundational work for robust multi-sensor alignment. More recently, Yuan has advanced visual place recognition (VPR), a key technology for robot localization and augmented reality. His 2024 study on deep homography estimation for VPR (17 citations) introduces a novel approach to improving hierarchical place recognition by refining geometric verification. In 2023, he proposed AANet (10 citations), an aggregation and alignment network that uses semi-hard positive sample mining to enhance the accuracy of two-stage VPR pipelines. Yuan’s contributions are shaping how autonomous systems perceive and navigate complex environments, bridging the gap between global retrieval and precise localization.