Jingen Liu

University of Michigan–Ann Arbor, SRI International

Papers

Total Citations

H-Index

About

Jingen Liu is a computer vision researcher whose work lies at the intersection of robotics, scene understanding, and motion analysis. His primary research focuses on enabling embodied agents—particularly mobile robots—to perceive and model their environments in real time using visual and motor cues. Liu’s most impactful contribution is his pioneering approach to indoor scene understanding, where he demonstrated that an agent can efficiently build a structural model of its surroundings by combining Bayesian filtering with motion cues derived from its own movement. This work, published in 2011 and garnering 65 citations, provides a foundation for real-time, autonomous navigation and spatial reasoning in cluttered indoor settings. Additionally, Liu has advanced motion segmentation techniques that leverage motor signals to separate foreground objects from background motion, a critical capability for robotic perception and interaction. His research also explores learning semantic features for visual recognition, contributing to broader applications in video indexing, surveillance, and human-machine interaction. Through these efforts, Liu has helped bridge the gap between low-level sensor data and high-level environmental understanding, making his work essential reading for researchers in robotics and active vision.