Feixiang Lu

Papers

Total Citations

H-Index

About

Feixiang Lu is a leading researcher at the intersection of 3D computer vision, robotics, and fine-grained object understanding. His work focuses on enabling intelligent agents to perceive and interact with articulated objects in dynamic environments. Lu’s major contributions include developing methods to parse part mobility from 3D data, such as his P^3-Net (2022, 5 citations), which learns explicit point correspondence from point cloud sequences to understand how object parts move. He also pioneered 3D Part Guided Image Editing (2020, 12 citations), a framework that uses 3D movable part models to enhance visual understanding for applications like autonomous driving. In InstanceFusion (2020, 8 citations), Lu introduced a real-time system combining deep learning with SLAM techniques to achieve instance-level 3D reconstruction of indoor scenes from a single RGBD camera. His work bridges the gap between geometric reconstruction and semantic understanding, with direct implications for robotics and augmented reality. Lu’s research is recognized for its practical impact, advancing how machines model the functional dynamics of objects—a critical step toward truly interactive AI.