Xidong Peng
Papers
2
Total Citations
7
H-Index
2
About
Xidong Peng is a researcher advancing the frontier of 3D visual grounding, with a focus on enabling machines to understand and interact with large-scale, dynamic real-world environments. His primary research areas lie at the intersection of computer vision, natural language processing, and multi-modal perception. Peng’s most notable contribution is the introduction of the WildRefer framework, which tackles the challenging task of 3D object localization in dynamic scenes using natural language descriptions and multi-modal visual data—specifically, 2D images and 3D LiDAR point clouds. This work, published in 2023 and 2024, has already garnered early recognition with 7 combined citations, signaling its importance in the field. By fully leveraging rich appearance and geometric cues, WildRefer bridges the gap between linguistic instructions and complex spatial reasoning, offering a robust solution for autonomous driving and robotics. Peng’s research is paving the way for more intuitive human-robot interaction and advanced scene understanding in unstructured environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2