Jeong Joon Park

University of Washington, University of Michigan–Ann Arbor

Papers

Total Citations

271

H-Index

About

Jeong Joon Park is a leading researcher at the intersection of computer graphics, 3D computer vision, and robotics, best known for pioneering neural implicit representations of 3D geometry. His landmark work, **DeepSDF**, introduced a continuous learned signed distance function that revolutionized how 3D shapes are represented, enabling high-fidelity reconstruction, rendering, and compression from partial or noisy data. This foundational paper has garnered over **259 citations**, becoming a cornerstone in the field and inspiring a wave of follow-up research on neural fields for geometry and appearance. Park’s contributions have directly influenced modern 3D content creation, autonomous navigation, and robotic manipulation. More recently, his work on **This&That** (2025) extends his impact into robot planning, where he combines language and gesture control to generate interpretable video instructions for complex tasks, bridging human-robot communication with visual reasoning. Through his innovative fusion of learning-based geometry and interactive robotics, Park continues to shape how machines perceive, represent, and act in the 3D world.