Andrea Tagliasacchi

École Polytechnique Fédérale de Lausanne, University of Toronto, University of Waterloo

Papers

Total Citations

787

H-Index

About

Andrea Tagliasacchi is a leading researcher at the intersection of computer vision, robotics, and 3D geometry processing. His work fundamentally advances how machines perceive, represent, and interact with the physical world. Tagliasacchi is best known for pioneering the "Sparse Iterative Closest Point" (SICP) algorithm (538 citations), a landmark contribution that dramatically improved the robustness and efficiency of 3D point cloud registration—a core task in robotics, autonomous navigation, and 3D scanning. More recently, he has been at the forefront of integrating deep learning with geometry for robotic manipulation. His highly influential work on "Neural Descriptor Fields" (NDFs) (138 citations) introduces SE(3)-equivariant object representations, enabling robots to generalize manipulation skills across different objects and poses with unprecedented accuracy. Tagliasacchi also made key contributions to multi-robot systems, developing collaborative dense scene reconstruction techniques (109 citations) that allow teams of robots to efficiently map unknown indoor environments. His research consistently bridges theoretical rigor with practical impact, shaping modern approaches to 3D perception and robotic autonomy.