Haram Kim
Papers
3
Total Citations
10
H-Index
2
About
Haram Kim is a robotics and computer vision researcher specializing in autonomous navigation and perception in dynamic environments. Their work centers on a particularly challenging problem in mobile robotics: detecting moving objects in real time so that robotic systems can navigate safely and accurately amid unpredictable surroundings. Kim's most significant contribution is the development of an occlusion accumulation framework for moving object detection, applied across multiple sensor modalities. Their 2020 work introduced this technique for RGB-D cameras, enabling visual odometry systems to distinguish moving objects from static backgrounds using depth information — a capability largely absent from prior color-image-based approaches. Building on this foundation, Kim extended the methodology to 3D LiDAR sensors in 2025, producing a lightweight, GPU-free solution capable of operating at real-time LiDAR frame rates without requiring environment-specific retraining — a notable advantage over prevailing deep learning methods. With a cumulative citation count of approximately 10 across their key publications, Kim's research addresses a critical bottleneck in robot autonomy. Their sensor-agnostic, computationally efficient approach makes their contributions particularly relevant for researchers and engineers working on real-world robotic deployment in unstructured, people-filled environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3