Pengbo Zhou

Papers

Total Citations

H-Index

About

Pengbo Zhou is a researcher specializing in 3D computer vision and point cloud processing, with a particular focus on point cloud registration—a critical technique for constructing high-precision 3D maps and enabling autonomous navigation, environmental perception, and robotic operation. His most notable contribution is the development of **PointTr**, a transformer-based architecture for low-overlap point cloud registration, which addresses the significant challenge of aligning point clouds in outdoor environments where overlapping regions are minimal. This work has garnered 13 citations since 2024, reflecting its timely relevance to the field. Zhou further advanced this line of research with **IOPCNet**, a novel method that classifies inner and outer points to achieve robust local-to-global registration even under extremely low overlap rates. His work is particularly impactful for applications in autonomous driving and robotics, where accurate 3D mapping under challenging conditions is essential. By tackling the long-standing problem of low-overlap registration, Zhou is helping to push the boundaries of what is possible in real-world 3D perception systems.