Hyuk‐Jae Lee
Papers
2
Total Citations
6
H-Index
2
About
Hyuk-Jae Lee is a leading researcher at the intersection of efficient deep learning and autonomous driving perception systems. His work focuses on two critical challenges: enabling real-time neural network inference on resource-constrained hardware, and improving the reliability of LiDAR-based environmental sensing. Lee’s most notable contribution is an accurate weight binarization scheme for CNN object detectors, specifically designed for YOLO-based systems used in robotics, drones, and autonomous driving. By introducing two scaling factors, his method dramatically reduces computational and memory requirements while preserving detection accuracy—a breakthrough for hardware implementation. This work has garnered significant attention, with 4 citations since 2020. Lee also addresses practical limitations of LiDAR sensors through his MLS framework, an MAE-aware sampling technique that leverages spatio-temporal information to improve point cloud acquisition in on-road environments. This innovation tackles the persistent challenges of low LiDAR resolution and high resource demands, earning 2 citations since 2021. Through these contributions, Lee is advancing the frontier of efficient, real-time perception for next-generation autonomous systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2