Alec Jacobson

Adobe Systems (United States), University of Toronto

Papers

Total Citations

H-Index

About

Alec Jacobson is a leading researcher in computer graphics, geometry processing, and robotics, with a focus on enabling machines to understand and manipulate 3D shapes. His most impactful work, "Neural Shape Mating: Self-Supervised Object Assembly with Adversarial Shape Priors" (2022, 34 citations), introduces a novel self-supervised framework for autonomous part assembly. Unlike traditional methods that rely on semantic part labels, Jacobson redefines assembly as a geometric mating problem, using adversarial shape priors to learn how to fit broken or disjoint pieces together without human annotation—a critical advance for robotic manipulation and 3D reconstruction. He also contributed to "Fast Updates for Least‐Squares Rotational Alignment" (2021, 7 citations), which optimizes the classic SVD-based rotation alignment, offering faster updates for real-time applications in graphics, vision, and simulation. By bridging geometric theory with practical algorithms, Jacobson’s work has influenced both academic research and industrial applications, particularly in automated assembly and shape understanding. His innovative, self-supervised approach marks a significant step toward more intelligent and autonomous robotic systems.