Leonidas J. Guibas

Leonidas John Guibas (; born 1949) is a Greek-American computer scientist and the Paul Pigott Professor of Computer Science (and, by courtesy, Electrical Engineering) at Stanford University, where he heads the Geometric Computation Group. His research spans computational geometry, computer graphics, computer vision, machine learning, and robotics, with contributions including foundational data structures, the earth mover's distance for image retrieval, Metropolis light transport, and the PointNet architecture for deep learning on point clouds.

Guibas is a member of the National Academy of Sciences, the National Academy of Engineering, and the American Academy of Arts and Sciences, and a Fellow of the ACM and the IEEE.

Education

Guibas was born and grew up in Athens, Greece. He received his B.S. and M.S. in mathematics from the California Institute of Technology in 1971, and his Ph.D. in computer science from Stanford University in 1976 under the supervision of Donald Knuth.

Career

After completing his doctorate, Guibas worked at Xerox PARC, DEC SRC, and MIT before joining the Stanford faculty in 1984. He was program chair for the ACM Symposium on Computational Geometry in 1996.

Research

Algorithms and data structures

Guibas's early work contributed several widely used data structures and algorithms in computational geometry. With Robert Sedgewick, he introduced red–black trees, a form of self-balancing binary search tree. Other contributions from this period include finger trees, fractional cascading, an optimal data structure for point location, the quad-edge data structure for representing planar subdivisions, and the Guibas–Stolfi algorithm for Delaunay triangulation. He also developed kinetic data structures for tracking objects in motion.

Computer graphics and vision

In computer graphics, Guibas co-authored work on Metropolis light transport, which enabled practical global illumination algorithms for photorealistic rendering. The EMD paper received the ICCV Helmholtz Prize in 2013, recognizing work with fundamental impact on computer vision.

Deep learning on point clouds and 3D geometry

More recently, Guibas's group has been a leader in applying deep learning to irregular geometric data such as point clouds and voxels. With Charles R. Qi, Hao Su, and others, he co-developed PointNet (2017), a neural network architecture that directly consumes raw point clouds for tasks including 3D object classification, part segmentation, and scene semantic parsing, without requiring conversion to voxel grids or image projections. The follow-up PointNet++ introduced hierarchical feature learning that captures local geometric structure at multiple scales. These architectures have been applied to problems in autonomous driving, robotics, and computational fluid dynamics.

Awards and honors

ACM Fellow (1999) ACM - AAAI Allen Newell Award (2007), "for his pioneering contributions in applying algorithms to a wide range of computer science disciplines" IEEE Fellow (2012) ICCV Helmholtz Prize (2013), for the earth mover's distance paper Member, National Academy of Engineering (2017) Member, American Academy of Arts and Sciences (2018) DoD Vannevar Bush Faculty Fellowship