Dongrui Liu

Papers

Total Citations

H-Index

About

Dongrui Liu is a researcher advancing the fields of 3D computer vision and robotics, with a primary focus on point cloud processing. His work addresses critical challenges in perception for autonomous systems, particularly in registration and recognition tasks. Liu’s most cited paper, "Point Cloud Registration using Representative Overlapping Points" (2021, 23 citations), tackles the fundamental problem of aligning 3D scans with partial overlap. He proposed a learning-based method that identifies representative overlapping points, moving beyond traditional correspondence-heavy approaches to achieve robust registration even under challenging conditions. This contribution is vital for applications like SLAM and 3D reconstruction. In his subsequent work, "A Robust and Reliable Point Cloud Recognition Network Under Rigid Transformation" (2022, 16 citations), Liu addressed a key vulnerability in modern point cloud networks: their lack of rotation robustness. He developed a model that maintains high recognition accuracy even under random rotations, a critical requirement for real-world deployment in industrial robotics and autonomous driving where sensor orientation varies. By targeting these practical limitations, Liu’s research directly improves the reliability of 3D perception systems, making them more resilient to real-world noise and transformations.