Bojun Ouyang

Papers

Total Citations

H-Index

About

Bojun Ouyang is a researcher working at the intersection of 3D computer vision and autonomous systems, with a focused contribution to the challenging problem of scene flow estimation on 3D point clouds. His most recognized work introduces an occlusion-guided self-supervised framework for estimating scene flow — the motion of points in 3D space between consecutive time frames — a capability fundamental to applications in virtual and augmented reality, robotics, and autonomous driving. What makes this contribution particularly significant is its emphasis on self-supervised learning, addressing a critical bottleneck in the field: the scarcity of real-world, non-simulated labeled data. By incorporating occlusion reasoning into the learning process, Ouyang's approach offers a more robust and practically deployable solution than supervised alternatives that rely on synthetic datasets. His work has garnered citations across the computer vision community, reflecting its relevance to researchers tackling perception and motion understanding in dynamic environments. While still an emerging voice in the field, Ouyang's research addresses a foundational challenge in geometric deep learning, and his self-supervised methodology positions him as a thoughtful contributor to the future of real-world 3D scene understanding.