Room “Sala Seminari” - Abacus Building (U14)
Robust Fitting in Computer Vision
Speaker
Prof. Luca Magri
Dipartimento di Elettronica, Informazione e Bioingegneria - Politecnico di Milano
Abstract
Robust model fitting is a cornerstone of computer vision, providing reliable estimation of geometric structures even in the presence of noise and outliers. This seminar revisits classical approaches such as Least Squares and RANSAC, and then moves beyond them to explore modern strategies for multi-model and multi-class fitting. These advanced techniques reinterpret robust estimation as a clustering problem, offering new perspectives on structure detection and data segmentation. Applications in feature matching, 3D reconstruction, and Scan-to-BIM demonstrate how robust fitting continues to play a central role in visual understanding and geometric reasoning.
Short Bio
Luca Magri is an Associate Professor at Politecnico di Milano. His research focuses on computer vision, 3D reconstruction, and robust model fitting, with particular interest in consensus-based and clustering approaches for multi-model estimation.
contact person for this Seminar: simone.melzi@unimib.it