Papers
5
Total Citations
93
H-Index
4
About
Hae-Gon Jeon is a leading researcher at the intersection of computer vision and robotics, with key contributions in autonomous navigation, 3D scene understanding, and computational imaging. His work on pedestrian trajectory prediction, particularly the "Disentangled Multi-Relational Graph Convolutional Network" (52 citations), advances how autonomous systems interpret social dynamics by modeling group-level interactions—a critical step for safe human-robot coexistence. In 3D sensing, Jeon developed the "Vari-Focal Light Field Camera" (19 citations), overcoming the depth-of-field limitations of conventional light field cameras to enable robust depth estimation for robotics and AR/VR applications. His self-supervised framework for off-road traversability estimation (10 citations) represents a paradigm shift, allowing mobile robots to learn terrain passability from past driving data without manual labels. Jeon also addresses critical gaps in disaster response through large-scale virtual datasets for egocentric localization (8 citations), and his early work on coded exposure motion deblurring for wheeled robots (2014) laid foundational techniques for real-time vision in dynamic environments. With over 90 total citations, Jeon’s research consistently bridges theoretical innovation and practical deployment, making him a pivotal figure in advancing perception systems for autonomous agents operating in unstructured, human-centered, and hazardous settings.
Research Focus
Key Achievements
Top Papers
- 1
- 2Vari-Focal Light Field Camera for Extended Depth of Field19 citations · 2021
- 3Self-Supervised 3D Traversability Estimation With Proxy Bank Guidance10 citations · 2023
- 4
- 5Motion deblurring using coded exposure for a wheeled mobile robot4 citations · 2014