Xiaojun Hou
Papers
4
Total Citations
14
H-Index
3
About
Xiaojun Hou is a rising researcher at the forefront of 3D scene understanding for autonomous systems and robotics. Their primary research areas encompass LiDAR panoptic segmentation and multi-modal 3D human tracking, with a strong focus on bridging the gap between state-of-the-art algorithms and real-world robotic deployment. Hou’s major contributions include pioneering novel frameworks like CenterLPS and PANet, which address the critical challenge of segmenting both semantic classes and object instances from LiDAR point clouds without relying on traditional, computationally expensive offset branches. This work, garnering early citations, is vital for enabling safe navigation in autonomous driving. In the domain of human-robot interaction, Hou has developed the Siamese Point-Video Transformer and a robotic-centric paradigm for 3D human tracking. These works tackle the fundamental issues of robustness in jittery, complex environments and the need for lightweight computational models suitable for on-robot deployment. With a growing body of work that directly addresses the practical limitations of current 3D perception, Hou is establishing a reputation for impactful, application-driven research that is essential for the next generation of intelligent, interactive robots.
Research Focus
Key Achievements
Top Papers
- 1CenterLPS: Segment Instances by Centers for LiDAR Panoptic Segmentation5 citations · 2023
- 2
- 3
- 4