Autocalibration via Rank-Constrained Estimation of the Absolute Quadric

Rank-constrained estimation of the absolute quadric

This page contains code/data/results for an autocalibration method that directly upgrades from a projective to metric reconstruction by estimating the dual absolute quadric. Main features of the algorithm are:

Enforces theoretical requirements such as positive semidefiniteness and rank degeneracy while estimating the absolute quadric.
Achieves the globally optimal solution for a reasonable objective function by constructing a series of convex LMI relaxations.
Complexity independent of number of views.

VRML Reconstructions:

VRML reconstructions for various datasets can be found here.
The zip file contains a README.pdf that details the contents and viewing instructions.

Reference:

M.K. Chandraker, S. Agarwal, F. Kahl, D. Nistér and D.J. Kriegman
Autocalibration via Rank-Constrained Estimation of the Absolute Quadric [PDF]
CVPR 2007, Minneapolis, Minnesota.

Last updated June 27, 2007.