Speaker
Jana Vráblíková
(Inria Centre at Université Côte d'Azur, Johannes Kepler University Linz)
Description
Real-time pose estimation is a fundamental challenge in computer vision. It involves finding a rotation and a translation that transforms one set of points into another. This problem is often posed as a non-linear least squares problem, which can be reduced to a homogeneous zero-dimensional polynomial system.
In this talk, we explain how to use resultant matrices to solve the problem in a closed-form, and we show how to exploit Sylvester forms to reduce its complexity.
Affiliation
Inria Centre at Université Côte d'Azur, Johannes Kepler University Linz
Author
Jana Vráblíková
(Inria Centre at Université Côte d'Azur, Johannes Kepler University Linz)
Co-authors
Ezio Malis
(Inria Centre at Université Côte d'Azur)
Laurent Busé
(Inria Centre at Université Côte d'Azur)