Gilhwan Kang
Papers
2
Total Citations
22
H-Index
2
About
Gilhwan Kang is making waves at the intersection of computer vision and robotics, with a sharp focus on perception in challenging, unstructured environments. His work tackles two fundamental problems: seeing clearly underwater and building robust maps in dynamic, real-world spaces. In his highly cited 2023 paper, “Joint-ID,” Kang introduced a novel transformer-based architecture that simultaneously performs image enhancement and depth estimation for degraded underwater scenes. By framing these tasks as a joint learning problem, his approach achieves superior results over traditional sequential methods, directly addressing the scattering and absorption that plague underwater imaging. More recently, with his 2025 work, “Uni-Mapper,” Kang has taken on the challenge of multi-modal LiDAR mapping. This framework provides a unified solution for creating consistent maps across different sensor types and dynamic environments—a critical capability for long-term, multi-session robot autonomy. While still early, this work signals his ambition to solve core infrastructure problems in field robotics. With a growing citation footprint, Kang is establishing himself as a researcher who builds practical, learning-based systems that push the boundaries of where robots can see and operate.
Research Focus
Key Achievements
Top Papers
- 1
- 2