Kaining Huang
Papers
2
Total Citations
19
H-Index
2
About
Kaining Huang is a rising researcher in computer vision and robotics, specializing in neural scene representations and 3D reconstruction under challenging environmental conditions. Their work bridges the gap between photorealistic rendering and real-world robotic perception, with a focus on enabling machines to see and understand scenes where traditional methods fail. Huang’s most notable contribution, **DarkGS** (2024, 13 citations), introduces a novel framework that combines neural illumination with 3D Gaussian splatting to reconstruct and relight scenes in low-light or dark environments—a critical capability for robotic exploration in subterranean, underwater, or night-time settings. This work demonstrates how robots can build consistent mental models of poorly lit spaces, mirroring human visual adaptability. Additionally, **RecGS** (2024, 6 citations) tackles the persistent problem of water caustics in seafloor imaging, using recurrent Gaussian splatting to remove dynamic light patterns without requiring annotated datasets or 2D filtering. By addressing these real-world visual challenges, Huang is advancing the frontier of robust, generalizable 3D perception for autonomous systems, with clear applications in marine biology, search-and-rescue, and planetary exploration. Their research is already shaping how robots interpret complex, unstructured environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2