Sidan Du

Papers

Total Citations

H-Index

About

Sidan Du is a leading researcher in 3D computer vision and autonomous systems, with a focus on urban scene understanding, LiDAR perception, and semantic mapping. Their pioneering work on classifying 3D urban street scenes from LiDAR point clouds—introducing a super-segment-based classification approach—has garnered 34 citations and laid critical groundwork for autonomous driving and robotics applications. Du further advanced the field with a real-time Truncated Signed Distance Field (TSDF)-based 3D semantic reconstruction method for LiDAR data, achieving both high-precision surface reconstruction and accurate semantic segmentation simultaneously. This work enables incremental mapping in dynamic environments, a key capability for self-driving cars and mobile robots. Du also contributed a comprehensive survey on online human action detection and anticipation in videos (34 citations), bridging temporal understanding with real-world deployment. Their research consistently addresses the gap between raw sensor data and actionable scene understanding, producing algorithms that are both theoretically sound and practically deployable. With a career focused on making machines perceive and navigate complex environments, Du’s contributions remain essential reading for researchers in autonomous navigation, 3D mapping, and intelligent transportation systems.