Dan Raviv

Papers

Total Citations

H-Index

About

Dan Raviv is a leading researcher in 3D computer vision and geometric deep learning, with a focus on self-supervised methods for understanding dynamic 3D scenes. His major contributions center on scene flow estimation—the task of tracking motion in 3D point clouds over time—a critical capability for applications in virtual and augmented reality, robotics, and autonomous driving. Raviv’s most-cited work, *Occlusion Guided Self-supervised Scene Flow Estimation on 3D Point Clouds* (2021), has garnered over 19 citations, addressing the persistent challenge of scarce labeled real-world data by introducing a self-supervised framework that leverages occlusion patterns to guide learning. This approach enables robust flow estimation without costly manual annotations, marking a significant advance in making geometric systems more practical and scalable. Raviv’s research is notable for its emphasis on real-world applicability, bridging the gap between simulated training and complex, unlabeled environments. His work continues to influence the development of efficient, data-driven solutions for 3D motion understanding, making him a key figure in the evolution of autonomous perception systems.