Dongki Jung

Papers

Total Citations

H-Index

About

Dongki Jung is a researcher specializing in computer vision and robotics, with a focus on 3D scene understanding and autonomous navigation. His key research areas include monocular depth estimation, visual localization, and novel-view rendering for ground robots. Jung's most notable contribution is **SelfTune** (2022, 4 citations), a self-supervised learning algorithm that resolves the scale ambiguity problem in monocular depth estimation by integrating monocular SLAM with proprioceptive sensors, enabling metric-scale depth prediction in real-world environments. He also developed **WayIL** (2024, 2 citations), a novel approach to indoor localization using abstract wayfinding maps, addressing the challenge of geometric discrepancies between maps and sensor data for robust robot localization. Most recently, Jung introduced **Mode-GS** (2024, 2 citations), a monocular depth-guided anchored 3D Gaussian splatting method that overcomes splat drift in ground-view scene rendering, significantly improving novel-view synthesis for robot trajectory datasets. His work bridges the gap between deep learning and classical geometric methods, advancing practical solutions for autonomous systems operating in complex indoor and outdoor environments.