Leonidas Guibas

Leonidas Guibas
Name	Leonidas Guibas
Fields	Computational geometry; Computer vision; Machine learning; Robotics
Workplaces	Stanford University; Xerox PARC; INRIA
Alma mater	Harvard University; Université Paris-Sud
Known for	Computational geometry algorithms; Shape analysis; Point cloud processing; Geometric inference

Leonidas Guibas Leonidas Guibas is a computer scientist known for foundational contributions to computational geometry, computer vision, machine learning, and robotics. His work spans theoretical algorithms, systems, and applied research connecting geometry with perception, graphics, and data analysis. He has held academic and industry positions and collaborated with researchers across Stanford University, Harvard University, Xerox PARC, and European research centers.

Early life and education

Guibas completed undergraduate and graduate studies that combined mathematical training and computer science exposure at prominent institutions. He earned degrees at Harvard University and pursued doctoral research and postdoctoral work involving collaborations with European laboratories such as INRIA and research centers linked to Université Paris-Sud and École Normale Supérieure. During his formative years he engaged with research communities associated with the ACM Symposium on Computational Geometry, European Conference on Computer Vision, and the IEEE Conference on Computer Vision and Pattern Recognition.

Academic career and positions

Guibas has been a faculty member at Stanford University where he held appointments bridging departments that include Computer Science Department, Stanford University and affiliated laboratories such as the Stanford Artificial Intelligence Laboratory and the Stanford Vision and Learning Lab. He has served in visiting and collaborative roles at industrial and academic institutions including Xerox PARC, INRIA, IBM Research, and exchanges with groups at Massachusetts Institute of Technology and University of California, Berkeley. Guibas has taught and advised students participating in venues like the SIGGRAPH community, the NeurIPS workshops, and the ICCV program committees. He has also been involved with organizing and contributing to initiatives sponsored by organizations such as the National Science Foundation and industry consortia including DARPA programs.

Research contributions and notable work

Guibas’s research portfolio integrates algorithmic theory and practical systems. In computational geometry he contributed to foundational topics including data structures for dynamic planar subdivisions, algorithms for nearest neighbor search, and combinatorial geometry analyses used in the Delaunay triangulation and Voronoi diagram literature. In computer vision and graphics he developed methods for shape matching, model registration, and surface reconstruction from point clouds, influencing pipelines used in 3D scanning, LiDAR processing, and photogrammetry. His work on shape representations, including spectral methods and probabilistic models, intersected with research on manifold learning and kernel methods in machine learning.

Guibas participated in early research on efficient motion planning and collision detection relevant to robotics and autonomous vehicles, linking geometric algorithms with practical systems used in robotic manipulation and perception stacks. He contributed to the development of point-set processing techniques, robust estimation methods such as variants of RANSAC, and probabilistic inference frameworks that were applied in projects associated with Google Research, Microsoft Research, and academic collaborations at ETH Zurich.

Interdisciplinary collaborations extended his influence into applications in medical imaging, cultural heritage digitization, and urban modeling. His group published methods for multi-view stereo, non-rigid registration, and learned descriptors for 3D shapes used in datasets maintained by communities around ShapeNet, ModelNet, and the Stanford 3D Scanning Repository.

Awards and honors

Guibas’s contributions have been recognized by awards and professional distinctions from major organizations. He has received fellowships and honors from bodies including the Association for Computing Machinery and the Institute of Electrical and Electronics Engineers for contributions to computational geometry and vision. His papers have been awarded best-paper recognition at conferences such as SIGGRAPH, CVPR, and SoCG (Symposium on Computational Geometry). He has been invited to deliver keynote and plenary talks at venues like NeurIPS, ICCV, and ECCV, and has served on editorial boards for journals such as the Journal of the ACM, IEEE Transactions on Pattern Analysis and Machine Intelligence, and Computational Geometry: Theory and Applications.

Selected publications and patents

Representative publications illustrate Guibas’s breadth across theory and applications: - Papers on dynamic geometric data structures and kinetic data structures presented at SoCG and published in proceedings associated with the ACM. - Work on surface reconstruction and point-set processing published in venues such as SIGGRAPH and CVPR and appearing in collections tied to the ACM Transactions on Graphics. - Contributions to shape analysis, spectral descriptors, and correspondence problems appearing in the IEEE Transactions on Visualization and Computer Graphics and conference proceedings of ECCV and ICCV. - Interdisciplinary papers on learning-based 3D representation and deep descriptors at NeurIPS and ICLR. - Patents and industrial collaborations produced during appointments with Xerox PARC and cooperative projects with industry teams at Google, Microsoft, and Apple addressing geometric algorithms for perception and mapping.

Selected monographs and influential articles have been cited widely by researchers working in computational topology, geometric modeling, computer graphics, and autonomous systems. His work is commonly included in curricula and textbooks that cover topics from algorithms in computational geometry to applied methods in computer vision.

Category:Computer scientists