Naiyan Wang

Papers

Total Citations

H-Index

About

Naiyan Wang is a prominent researcher whose work sits at the intersection of autonomous driving, robotics, and computer vision. His research focuses on critical perception challenges including LiDAR-based localization, 3D object tracking, and implicit shape reconstruction — all foundational technologies for reliable autonomous systems operating in complex real-world environments. Wang's most recognized contribution is **DMLO: Deep Matching LiDAR Odometry** (2020), which has accumulated over 51 citations and addresses one of robotics' core challenges: accurately estimating a vehicle's trajectory from LiDAR sensor data. Rather than relying solely on conventional local iterative approaches, DMLO leverages deep learning-based feature matching for more robust global registration, significantly improving performance on noisy, real-world data. This work represents a meaningful departure from traditional odometry pipelines. His more recent research on implicit object tracking and shape reconstruction (2022) pushes toward online adaptability, tackling the difficult problem of generalizing 3D reconstruction methods to cluttered, unconstrained scenes — a critical gap for deployment in practical autonomous driving systems. Wang's body of work demonstrates a consistent commitment to bridging the gap between theoretical computer vision and the robustness demands of real-world autonomous systems, making him a noteworthy contributor to the field.